Head-to-Head Controlled Trials
|
Bahk (2005) South Korea (Poor) | Multicenter (8 sites), open-label
RCT University-based hospitals, tertiary care unit, and chronic
mental health institute | DSM-IV bipolar I disorder with current manic
episode and requirement for antipsychotic treatment; age 18 to 65 y; minimum
score on Young Mania Rating Scale (YMRS) of 20; medicosurgically stable | Topiramate + Risperidone vs.
Divalproex + Risperidone, flexibly dosed for 6
wk Recommended starting dose (and titration rate every 2 to 5 d):
Topiramate 50 mg/d (rate: 25 to 50 mg/d); Divalproex 750 mg/d (rate: 250 to
500 mg/d); Risperidone 0.5 to 2 mg/d (clinician’s judgment) | 3-d washout of prior medications | Oral lorazepam < / = 4 mg/d;
injectable lorazepam except within 24 h before completing rating scales;
antiparkinsonian drugs | YMRS, Clinical Global Impression (CGI),
Simpson-Angus Rating Scale (SARS, neurologic adverse events) at baseline, wk
1, wk 3, and wk 6 / endpoint; reduction in YMRS and CGI scores of > /
= 50% at end point vs. baseline; vital signs and
adverse events at all assessment periods; ECG and blood tests at baseline
and end point. Remission defined as YMRS < / =
12. | Topiramate vs. Divalproex (each combined with
Risperidone) Age, mean, y: 37.5 vs. 37.6 Male, n
(%): 15 (56.6%) vs. 22
(53.7%) Ethnicity: Not reported | YMRS: 35.2 vs. 33.9 CGI-s: 5.3 vs.
5.5 SARS: 0.2 vs. 0.5 Age at onset, y: 29.3 vs.
38.8 Body mass index (BMI), kg/m2 : 24.1 vs.
24.6 Weight, kg: 65.4 vs. 67.3
Drug use within 1 y
prior to study, n (% of total patients): --Mood
stabilizer: 44 (59.5%) --Antipsychotic: 14
(18.9%) --Antianxiety: 56
(75.7%) --Antidepressant: 8
(10.8%)
Most common drug used within 1 y prior to
study --Mood stabilizer, lithium, n: 15 vs.
17 --Antipsychotic, olanzapine: 2 vs. 4 --Anxiolytic,
alprazolam: 17 vs. 21 --Antidepressant, paroxetine: 2 vs. 3 | 81 screened / number eligible not reported / 74
enrolled and randomized | 10 withdrew / 3 lost to follow-up / 74
analyzed | Topiramate (N = 33) vs.
Divalproex (N = 41) Doses, mean,
mg/d --Mood stabilizer: 220.6 vs. 908.3 --Risperidone: 3.4 vs.
3.3 (NSD) --Lorazepam: 1.8 vs. 1.5 (NSD) --Benztropine: 1.4
vs. 1.8 (NSD)
Absolute (Relative) decrease in
scores --YMRS: 23.9 (67.9%) vs.
21.6 (63.7%) (NSD) --CGI: 3.0 (56.6%)
vs. 3.2 (58.2%) (NSD) | Responder rates (Patients with > /
= 50% reduction), n (%) --YMRS:
25 (75.8%) vs. 29 (70.7%) (NSD) --CGI-s: 24
(72.7%) vs. 30 (73.2%) (NSD) | Patients entering remission (YMRS < /
= 12), n (%): 21 (63.6%) vs. 25
(61.0%) (NSD) | | Monitoring | Topiramate (N = 33) vs. Divalproex (N
= 44) (each in combination with Risperidone) AEs reported
in > / = 10% of patients in either treatment
group, n (%) --Dizziness: 7 (21.2%) vs. 0
(0%) --Headache: 6 (18.2%) vs. 2
(4.9%) --Nausea: 4 (12.1%) vs. 5
(2.4%) --Paresthesia: 3 (6.8%) vs. 0
(0%) --Sedation: 1 (3.0%) vs. 8
(19.5%) --Concentration difficulty: 1 (3.0%)
vs. 6 (14.6%) Other AEs: --Extrapyramidal symptom:
9 (27.3%) vs. 13 (31.7%) --Increased alanine
aminotransferase (ALT): 1 (3.0%) vs. 2
(4.5%) SARS score, mean change from baseline to end
point: Values not reported (NSD) Patients showing weight change at
end point, n (%) --Weight loss in topiramate group: 15
(45.5%) --Weight gain in divalproex group: 30
(73.2%) Mean change from baseline to end
point --Weight, kg (%): −0.25
(0.5%) vs. 2.25 (3.6%) (p <
0.0001) --BMI, kg/m2 (%):−0.1
(0.4%) vs. 0.75 (3.3%) (p < 0.0001) | Total withdrawals, n: 5 vs. 8
(NSD) Withdrawals due to AEs: Not reported by group | AE rates reflect combination therapy; no
monotherapy control group for comparison. In post hoc analyses, no
correlation was found between weight loss with topiramate and topiramate
dose, initial weight, BMI, and gender. Possible observer biases due to
multicenter design. Possible carryover effects of prior treatments due to
relatively short washout period. |
Frye, 2000 U.S. (Fair) | DB RCT with two crossovers Single center, National
Institute of Mental Health (NIMH) Clinical Research Unit, inpatient setting
Extension of this trial by Obrocea, 2002 | Not explicitly listed. Refractory bipolar and
unipolar affective illness confirmed by the Structured Clinical Interview
for DSM-IV Axis I Disorders (version 2.0), hospitalized in NIMH Clinical
Research Unit. Illness did not respond to conventional agents | Lamotrigine (titrated from 25 to 500 mg/d over 5
to 6 wk, faster than current product labeling at the time of the study)
vs. Gabapentin (titrated from 900 to 4800 mg/d) vs. Placebo for 6
wk | 1-wk washout before crossover: taper old drug,
titrate new drug | Levothyroxine; diuretic; triiodothyronine,
clonazepam | Clinical Global Impression scale modified for
bipolar illness (CGI-BP), timing not reported. CGI-BP best estimate
rating determined after completion of each 6-wk treatment phase | Age, mean (SD), y: 39.2 (9.4) Male /
Female: 42% / 58% Ethnicity not reported | Bipolar I 36% Bipolar II
45% Unipolar 19% Rapid cycling
92% Nonrapid cycling 8% Prior
treatment (N Refractory/N Exposed, %): Lithium 28/28
(100%) Valproic acid 21/26
(81%) Carbamazepine 14/20 (70%) | Number screened not reported / 38 eligible / 38
enrolled / 38 randomized | 4 withdrawn / 0 lost to follow-up / 31 analyzed (3
not evaluable in all three phases and excluded from Cochran’s
Q analysis) | Lamotrigine vs. Gabapentin
vs. Placebo Responders (score of much or very much improved on
Clinical Global Impressions Scale for Bipolar Illness) after 6 wk on each
treatment: Mania, 44% vs. 20% vs.
32% (NSD) Depression, 45% vs. 26%
vs. 19% (NSD) Overall, 52% vs.
26% vs. 23% (p = 0.031; post hoc Q
differences: p = 0.011 for lamotrigine vs. gabapentin; p
= 0.022 for lamotrigine vs. placebo; p =
0.700 for gabapentin vs. placebo) | Lamotrigine vs. Gabapentin Mean change in Hamilton
Rating Scale score for Depression (HAM D) from baseline to 6 wk:
−6.1 vs. 1.6 (placebo result not reported) Calculated
difference between mean changes: −7.7 Changes from
baseline to 6 wk in Speilberger State Anxiety Scale, Young Mania Rating
Scale (YMRS), and Brief Psychiatric Rating Scale (BPRS): NSD (data not
reported). | | | Not reported | Lamotrigine: Rash developed post-study in wk 15
during continuation treatment, progressed to toxic epidermal necrolysis;
patient required hospitalization in an intensive care burn unit and fully
recovered.
Lamotrigine vs. Gabapentin vs. Placebo (N
= 31) Weight change, mean (SD): −0.96 (3.11) vs.
1.83 (5.04) vs. −0.40 (2.97) kg (p = 0.024; for
lamotrigine vs. gabapentin, p = 0.021; p > 0.05 for
lamotrigine vs. placebo and for gabapentin vs. placebo) Common
adverse effects: --Ataxia 3% vs. 10% vs.
0% --Diarrhea 6% vs. 6% vs.
13% --Diplopia 0% vs. 10% vs.
3% --Fatigue 0% vs. 10% vs.
3% --Headache 3% vs. 13% vs.
13% --Rash 3% vs 0% vs
0% | Lamotrigine vs. gabapentin Total Withdrawals :
3/38 (7.9%) vs. 1/38 (2.6%); 1 additional patient
(treatment group not reported) withdrew due to nonresponse. Withdrawals due
to adverse event: 3/38 (7.9%) vs. 1/38 (2.6%) (no
statistical analysis) The gabapentin patient was the same as one of
the lamotrigine patients; patient withdrew after developing edema on both
drugs. Types of withdrawals due to adverse event: rash, edema on
lamotrigine; edema on gabapentin. | Heterogeneous study population. Lamotrigine dose
titrated at faster than currently recommended rates. |
Obrocea,
2002 U.S. (Fair) Same trial as Frye 2000 | DB RCT with two crossovers; extension of Frye,
2000; analyzed subgroup response predictors Single center, National
Institute of Mental Health (NIMH) Clinical Research Unit, inpatient
setting | Not explicitly listed. Refractory bipolar and
unipolar affective illness confirmed by the Structured Clinical Interview
for DSM-IV Axis I Disorders (version 2.0), hospitalized in NIMH Clinical
Research Unit. Illness did not respond to conventional agents | Lamotrigine (titrated from 25 to 500 mg/d over 5
to 6 wk, faster than current product labeling at the time of the study) vs.
Gabapentin (titrated from 900 to 4800 mg/d) vs. Placebo for 6 wk | 1-wk washout before crossover: taper old drug,
titrate new drug | Levothyroxine; diuretic; triiodothyronine,
clonazepam | Clinical Global Impression scale modified for
bipolar illness (CGI-BP), timing not reported. CGI-BP included
Hamilton Depression Rating Scale (HAM-D); clinician and self
prospective Life Chart Method (LCM), Young Mania Rating Scale
(YMRS); Spielberger State Anxiety Scale; and Bunney-Hamburg ratings
of depression and mania | N = 45 Age, mean (SD), y: 39.2
+/− 10.5 Male / Female: 40% /
60% Ethnicity not reported | Bipolar I 33% Bipolar II
44% Unipolar 22% Rapid cycling
74% Exposed, calculated %): --Lithium
34/40 (85.0%) --Valproate 23/35
(65.7%) --Carbamazepine 15/25
(60.0%) Hospitalizations, mean (SD) --Mania,
bipolar: 0.9 (1.8) --Mania, unipolar: 0.0 (0.0) --Depression,
bipolar: 3.6 (3.5) --Depression, unipolar: 2.6 (2.8) | Numbers screened and eligible not reported / 45
enrolled / 45 (?)randomized | Numbers withdrawn and lost to follow-up not
reported / 38 to 40 analyzed depending on treatment group | Lamotrigine vs. Gabapentin vs.
Placebo Responder rate for CGI-BP much or very much
improved All exposed to given drug: 20/39 (51%) vs. 11/40
(28%) vs. 8/38 (21%) (no statistical
analysis) Exposed to all 3 phases of protocol (N = 36):
53% vs. 28% vs. 22% (p =
0.01; NSD for gabapentin vs. placebo)
CGI ratings for
depression showed a similar pattern (p = 0.03) | Predictors of response to lamotrigine (using
CGI-BP overall degrees of improvement or deterioration): --Diagnosis
of bipolar illness (r = − 0.32; p =
0.49) --Male gender (r = 0.37; p = 0.022)
--Exposure to fewer prior medication trials (r =
−0.40; p = 0.015) --History of fewer prior
hospitalizations for depression (r = −0.32; p
= 0.050)
Factors influencing amount of variance
explained by the predictors (stepwise linear regression): --Number of
prior medication trials (Beta coefficient = −0.369;
p = 0.018) --Gender (Beta coefficient =
0.357; p = 0.021) Similar beta coefficients suggested
that these variables had equal importance in predicting lamotrigine
response. Adjusted R
2 showed that these variables explained 24% of the
variance of CGI response. | Possible predictors of response to
gabapentin --Duration of illness inversely correlated with response
(r = −0.35; p = 0.028) --Weight
at baseline inversely correlated with response (r =
−0.44; p = 0.006)
Stepwise linear
regression analysis: --Age (Beta coefficient − 0.492; p
= 0.001) --Weight (Beta coefficient =
−0.493; p = 0.001) Similar beta coefficients
suggested that these variables were equally important in predicting response
to gabapentin. Adjusted R
2 showed that these variables explained 37% of the
variance of CGI response. | | Not reported | Not reported | Not reported | A post hoc test was used for specific paired
comparisons. |
Suppes, 2007 U.S | SB RCT, open-label, outpatient | DSM-IV diagnosis of bipolar I, II or not otherwise
specified; no medications or a stable medication regimen for at least 1m
prior to study entry; aged 18–65y; sodium serum levels between
134–146 mEq/L; currently experiencing hypomania (YMRS
≥ 12), confirmed on at least 2 occausions prior to
randomization; no substance abuse/dependence within the past month; non-
pregnant and not nursing; no hypersensitivity to oxcarbazepine or
carbamazepine; no severe liver disease or dysfunction; no hyponatremia; no
suspecion of chronic infectious disease | Oxcarbazepine 300 mg/d (increased to maximum 2400
mg/d, target dose 1200 mg/d, mean 1350 mg/d) vs Divalproex 500 mg/d
(increased to minimum blood level of 50 mg/mL, mean 1167 mg/d) for 8w | Not reported | Lorazepam 10 mg allowed for acute agitation; no
other changes in ongoing medications were
permitted
Oxcarbazepine vs. Divaproex Lithium:
1(6.7%) vs. 1(6.7%) Anticonvulsants:
0(0%) vs. 2(13.3%) Antidepressants:
6(40.0%) vs. 3(20.0%) Antipsychotics:
1(6.7%) vs. 2(13.3%) Psychostimulants:
0(0%) vs. 1(6.7%) Anti-anxiety drugs:
1(6.7%) vs. 3(20.0%)
Number of types
of concomitant medications across groups found to be non- significant
(p=0.56) | Clinical symptoms rated using YMRS, the IDS C, and
the Clinical Global Impressions scale for use in bipolar disorder | Oxcarbazepine vs. Divaproex Age: 30.1(8.0)
vs. 36.9(9.9); p<0.05 Male (%): 6(40%)
vs. 6(40%) Ethnicity: White: 14(93.3%)
vs. 10(66.7%) Hispanic: 1(6.7%) vs.
4(26.7%) Other: 0(0%) vs.
1(6.7%) | Oxcarbazepine vs. Divaproex Bipolar I:
4(26.7%) vs. 8(53.3%) Bipolar II:
7(46.7%) vs. 4(26.7%) Bipolar NOS:
4(26.7%) vs. 3(20.0%)
Weight
(lbs): 84.99(26.2) vs. 84.09(26.5); p=0.91
YMRS:
22.07(5.86) vs. 20.53(6.02)
CGI1c: 4.08(0.86) vs.
4.00(0.68) | Number screened not reported / 30 eligible / 30
enrolled / 30 randomized | 13 withdrawn / 6 lost to follow-up / 17
analyzed | Oxcarbazepine vs. Divaproex
YMRS,
mean change: −63.8% vs. −
79.0%; p<0.001 vs baseline; between-group,
p=0.95
IDS-C, mean change:
−48.7% vs. − 19.7%;
p=0.82
CGI1c: 2.00(1.63) vs. 2.75(1.75);
between-group, p=0.37 | | | | AEs reported and assessed at every visit | Oxcarbazepine vs. Divaproex Side effects,
median: 2.0 vs. 3.0;
between-group, p=0.29
Most common side
effects were drowsiness or sedation in both groups; other common side
effects were dizziness or lightheadedness, blurred vision, increased thirst
and headaches in the oxcarbazepine group; tiredness, decreased appetite, and
weight gain in the divaproex grop
No patients developed
hyponatremia | Oxcarbazepine vs. Divaproex Total
withdrawals: 13(53.3%) Withdrawals due to AEs:
1(3.3%)
Other reasons for withdrawal inlcude
worsening mood symptoms, lack of improvement, and personal reasons | Clinical symptoms were evaluated weekly for 4w,
then biweekly for 4w for a total of 8w
16 patients used
monotherapy, 14 as add-on treatment; detailed concomitant medical treatment
described in text
Results from a tw-wave ANOVA with
one between- subjects factor (group) and one within-subjects factor showed
non-signficant effects for groups (p=0.81), visits
(p=0.25) and group/visit interaction
(p=0.31)
Rater blind to treatment assignments
completed all ratings |
Vasudev, 2000 India (Poor) | SB RCT Single-center, psychiatric inpatient
setting | Bipolar disorder (DSM-III-R), Young Mania Rating
Scale (YMRS) >/= 20 | Carbamazepine titrated, 800 to 1600
mg/d Sodium valproate titrated, 800 to 2200 mg/d for 4 wk | None | Diazepam, promethazine | YMRS weekly from day 0 to 28 for valproate and at
days 0 and 10 then weekly to day 31 for carbamazepine (different schedules
were used because a therapeutic dose of carbamazepine was reached at day
3) | Not reported | Not reported | Numbers screened and eligible not reported/30
enrolled/30 randomized | 6 (20.0%) withdrew/lost to follow-up
NR 30 analyzed | Carbamazepine vs. Valproate YMRS total
scores, mean change from baseline to day 28 (Primary Efficacy Measure; last
observation carried forward): 20.8 vs. 32.8 (calculated difference:
−12; p = 0.023) | Weekly analysis of change in YMRS
scores Decrease in scores on YMRS --Week 1: Data not reported
(NSD) --Week 2 and on: Valproate superior to carbamazepine (data not
reported; p = 0.04)
Response
analysis > 50% decrease in YMRS total score from
baseline to end point: 8/15 (53.3%) vs. 11/15
(73.3%) (NSD)
YMRS individual
items Valproate showed a numerically greater mean improvement vs.
carbamazepine except for sleep. | Required rescue medication Week 1: NSD
(data not reported) Week 2: 12/15 (80.0%) vs. 4/15
(26.7%) (p = 0.003)
Average dose of
rescue medication required, mg/d (estimated from Fig. 1 of article) Week
1 --Diazepam: 16 vs. 10 --Promethazine: 72 vs. 55 Week
2 --Diazepam: 8 vs. 1 --Promethazine: 40 vs. 10 | | Not reported | Carbamazepine vs.
Valproate
Experienced adverse events: 67% vs.
17%
Adverse events more common on
carbamazepine --Nausea/vomiting: 58.3% vs.
16.7% (p = 0.035) --Dizziness:
58.3% vs. 8.3% (p =
009) --Lethargy: 41.6% vs. 8.3% (no
statistical analysis) --Ataxia/Tremors: 25% vs.
8.3% (no statistical analysis) --Rash: 8.3%
vs. 0.0% (no statistical analysis) --Increased liver
enzymes: 8.3% vs. 8.3% --Hematologic
abnormalities: 0% vs. 0% | Total withdrawals: 3 vs. 3 Withdrawals due
to adverse events: 1 vs. 0 (withdrawal on carbamazepine due to severe
vomiting was temporary) | Unclear if care provider was the unblinded dosing
psychiatrist. Medications were apparently not identical. Titration phases to
therapeutic dose were of different durations (3 vs. 0 d on carbamazepine vs.
valproate, respectively) and may have favored faster onset of effect with
valproate, since a therapeutic (loading) dose of 20 mg/kg could be given on
the first day. Drug exposure time and end point differed between treatment
groups: 31 vs. 28 d. |
Active-Controlled Trials
|
Bowden, 1994 U.S. | Multicenter, randomized, double-blind,
parallel-group, placebo-controlled study Inpatient setting | Men and women 18–65y who met Research
Diagnostic Criteria for manic disorder based on the structured interview and
rating scale of the Schedule for Affective Disorders and Schizophrenia; YMRS
≥ 14 | Divalprox sodium 750 mg/d tid (increased to 1000
mg/d at 3d) vs. Lithium 900 mg/d tid (increased to 1200 mg/d at 3d) vs.
Placebo | 3 to 21 day washout period based on half-life of
the psychoactive drug taken on admission | Patients terminated from study if they required
psychoactive medication; protocol allowed use of adjunctive chloral hydrate
(maximum 4 g/d) or lorazepam (maximum 2 mg/d) as need for control of
agitation, irritability, restlessness, insomnia and hostile behaviors;
medications not permitting within 8h of assessment; neuroleptic drugs not
allowed NSD observed among groups for antimania medication use | SADS, Research Diagnostic criteria, SADS-C, GAS on
the first and last day of study; Affective Disorder Rating Scale (ADRS)
given on days 3, 5, 7, 9, 12, 18, and 20; SADS-C and GAS administered on
days 5, 10, 15, and 21; YMRS | Divalproex vs. Lithium vs. Placebo Age:
40.4(12.8) vs. 39.1(11.2) vs. 39.0(10.0) Male
(%):36(52%) vs. 26(72%) vs.
42(57%) Ethnicity: White: 50(73%) vs.
24(66%) vs. 53(74%) Black:
12(17%) vs. 6(17%) vs.
14(19%) Other: 7(10%) vs. 6(17%)
vs. 7(9%) | Divalproex vs. Lithium vs. Placebo Duration of
illness (y)_: 18.0(12.4) vs. 16.1(11.0) vs. 18.0(10.4) Any major mood
episodes: 11(19%) vs. 1(4%) vs.
6(10%) Mania episode: 8(14%) vs.
0(0%) vs. 0(0%); p<0.05
Groups
comparable except 8 patients with > 4 episodes were in the divalproex
group
Prior lithium treatment: 54(78%) vs.
31(86%) vs. 61(82%) Effective, tolerated:
22(41%) vs. 16(52%) vs.
19(31%) Effective, not tolerated: 7(13%) vs.
0(0%) vs. 6(10%) Ineffective, tolerated:
19(35%) vs. 11(35%) vs.
31(51%) Ineffective, not tolerated: 6(11%)
vs. 4(13%) vs. 5(8%) | Number screened not reported/number eligible not
reported/179 enrollled/179 randomized | 82 withdrew/lost to follow-up not reported/176
analyzed (ITT population) | Divalproex vs. Lithium vs.
Placebo
GAS, mean change: 7.6 vs. NR vs. 3.8;
p=0.06
ADRS, mean change: Mania: −4.9
vs. −5.9 vs. −0.2 Elation/grandosity:
−2.6 vs. NR vs. − 0.7 Psychosis:
−2.7 vs. NR vs. 0.6
> 50%
improvement, Manic Syndrome subscore: 48% vs. 49%
vs. 25%; p=0.004, divalproex vs. placebo;
p=0.025, lithium vs. placebo | Divalproex patients had greater improvement on
Manic Syndrome subscale (day 5), MRS (day 10) and the Behavior and Ideation
subscale score than placebo Divalproex had greater improvement in
elevated modd, less need for sleep, excessive activity and motor
hyperactivity than placebo; greater improvement in excessive activity and
motor hyperactivity than lithium group | Divalproex vs. Lithium vs.
Placebo
Of 142 patients with known lithium responsiveness
prior to the study:
Responders: MRS, mean change:
−7.4 vs. −15.3 vs.
−4.0
Non-responders: MRS, mean change:
−10.8 vs. −1.0 vs. −3.2 | | Monitored | Divalproex vs. Lithium vs. Placebo AEs:
58(84%) vs. 33(92%) vs.
58(78%) Asthenia: 9(13%) vs.
7(19%) vs. 7(9%) Constipation:
7(10%) vs. 6(17%) vs.
5(7%) Diarrhea: 8(12%) vs. 5(14%)
vs. 13(18%) Dizziness: 11(16%) vs.
3(8%) vs. 4(5%) Fever: 1(1%) vs.
5(14%) vs. 3(4%); p<0.05, lithium vs.
divalproex Headache: 15(22%) vs. 14(39%) vs.
24(32%) Nausea: 16(23%) vs.
11(31%) vs. 11(15%) Pain: 13(19%)
vs. 1(3%) vs. 15(20%); p<0.05,
lithium/divalproex vs. placebo Somnolence: 13(16%) vs.
7(19%) vs. 11(15%) Twitching:
2(3%) vs. 3(8%) vs. 0(0%);
p<0.05, lithium vs. placebo Vomiting: 10(14%) vs.
9(25%) vs. 3(4%); p<0.05, divalproex vs.
lithium/placebo
No AEs related to bleeding or bruising
occurred in any patient with reduction in platelet count (platelet count,
mean change: 77x10^9/L divalproex only); hepatic function did not change or
improve with divalproex
Lithium patients had increased
platelet, WBC, and neutrophil counts
No other clinically significant changes observed in any group | Divalproex vs. Lithium vs. Placebo Total
withdrawals: 33/69(48%) vs. 22/36(61%) vs.
47/74(64%) Withdrawals due to AEs: Not reported (rate of
intolerance to treatment higher in lithium group) | Medication adjustments made by an unblinded
physician throughout the study Patients terminated from study if they
had a reducation > 50% from baseline in SADS-C Mania
Rating score; no SADS-C Mania rating score > 2; and GAS score
> 70 |
Maina, 2006 Italy | RCT, open-label, add-on therapy, single-center
(Mood and Anxiety Disorders Unit of the University of Turin) | Patients with a diagnosis of bipolar disorder,
manic or hypomanic episodel YMRS ≥ 16; Hamitlton Depression
Rating Scale ≤ 7; > 1y lithium
treatment
Patients excluded if administered other concurrent
drugs (except benzodiazepines) during index manic or hypomanic episode | Valproate 500-1500 mg/d (mean dose 972 mg/d) vs.
Olanzapine 7.5–15.0 mg/d (mean dose 11.25) add-on to lithium for
8w | Not reported | Benzodiazepines; lithium dosages maintained
unchanged throughout study; no other concmitant medications allowed during
study | YMRS, CGI-S and CGI-I administered weekly
throughout study | Valproate vs. Olanzapine Age: 48.6(12.8)
vs. 45(12.4); p=0.528 Male
(%):5(56.6%) vs. 7(58.3%);
p=0.899 Ethnicity: Not reported | Valproate vs. Olanzapine Bipolar I:
4(44.4%) vs. 5(41.7%);
p=0.899
Duration of illness (y): 18.0(7.45) vs.
16.4(10.7); p=0.236
Lithium blood levels
(mmol/L): 0.78(0.13) vs. 0.76(0.12); p=0.679 | Number screened not reported/number eligible not
reported/21 enrolled/21 randomized | 0 withdrawn/0 lost to follow-up/21 anaylzed | Valproate vs. Olanzapine Mean change, YMRS:
−17.67(6.89) vs. −20.08(5.64);
p=0.367 Mean change, CGI-S: −2.56(0.88) vs.
−3.08(0.66); p=0.100 CGI-I: 1.44(0.52) vs.
1.33(0.49); p=0.625
Both groups showed
significant reduction (p<0.001)
YMRS scores associated
with type of drug (p=0.014); type of BPD (p=0.389)
and duration of illness (p=0.836) had not effect on predicting
YMRS reduction | Valproate vs. Olanzapine Responders:
6(66.7%) vs. 10(83.3%);
p=0.375
Response at week 4 significant
between-group, p=0.030 in favor of
Olanzapine
Remission: 5(55.6%) vs.
7(58.3%); p=0.899 | | | AEs reported at every time pont by direct
interview with patinent; no administration of specific scale | Valproate vs. Olanzapine Somnolence:
1(11.1%) vs. 3(25.0%) Tremor:
2(22.2%) vs. 2(16.7%) Weight gain:
1(11.1%) vs. 1(8.3%) Headache:
1(11.1%) vs. 1(8.3%) | Total withdrawals:
0(0%) Withdrawals due to AEs: 0(0%) | MMRM showed difference between groups disappeared
over time for the CGI-S outcome |
Ichim, 2000 South Africa | DB, RCT, pilot study, inpatient | Male and female inpatients, aged
18–65y; admitted with an acute manic episode. All must meet
DSM-IV criteria for manic phase of BPD
Patients excluded if
they had abnormal liver function, thyroid function, or hematological
findings; other acute medical disorder or medical disorder requiring
frequent changes in medication; received treatment with neuroleptic deport
preparation within the last month or fluoxetine within past 5w; or had drug
or alcohol abuse | Lamotrigine (titrated from 25 mg/d at 1w to 50
mg/d at 2w and finally, 100 mg/d at 3w) vs. Lithium 400mg bid | 1d; a longer wash-out period not practical due to
difficulty of withholding treatment from disturbed patients | Lorazepam 4–12 mg/d as rescue
medication allowed to control aggression; any exisiting psychotropic
medication discontinued for > 1d before entering study | Psychiatric condition measured using MRS, Brief
Psychiatric Rating Scale (BPRS), CGI, and the Global Assessment Functioning
scale every week for 4w | Lamotrigine vs. Lithium Age: 33.6 vs.
31.9 Male (%):8(53.3%) vs.
8(53.3%) Ethnicity: Not reported | Lamotrigine vs. Lithium Duration of episode
(d): 13.3 vs. 19.3; p=0.048; p=0.076 with outlier
removed Depressive episodes: 0.7 vs. 0.7 Previous admissions:
1.8 vs. 2.9
BPRS: 52.8 vs. 46.8 MRS: 34.4 vs.
31.6 CGI-S: 4.93 vs. 4.67; NSD GAF: 44.7 vs. 45.5 | Number screened not reported/number eligible not
reported/30 enrolled/30 randomized | 5 withdrawn/lost to follow-up not reported/20
analyzed | Lamotrigine vs. Lithium BPRS: 30.2
(p=0.0002) vs. 28.2(0.0005); NSD between groups MRS: 14.3
(p=0.0002) vs. 13.2(p=0.0005); NSD between
groups CGI-S: 2.77(p=0.0002) vs.
2.83(p=0.0005); NSD between groups CGI-I: Both groups
showed improvement, NSD between groups GAF: 55.7
(p=0.001) vs. 56.2 (p=0.002); NSD between
groups | Lamotrigine vs. Lithium Response rate
(≥ 50% reduction in MRS): 8(53.3%) vs.
9(60.0%) Response rate (≤ 50%
reduction in BPRS): 7(46.7%) vs.
4(26.7%) Response rate
(CGI-S=1–2): 7(46.7%) vs.
4(26.7%) | Lamotrigine vs. Lithium Rescue medication,
mean dose (mg/d): 2.65 vs. 2.66; NSD between groups | | BP measured at each visit | No significant AEs noted in either group; no
rashes reported in Lamotrigine group | Lamotrigine vs. Lithium Total withdrawals:
2/15(13.3) vs. 3/15(20%) Withdrawals due to AEs: Not
reported | Selection bias; absence of placebo group; low
dosages to minimize AEs |
Okuma, 1979 Japan | DB, RCT, multi-center Inpatient and outpaient
setting | Untreated patients diagnosed as endogenous manics
(i.e., manic-depressive psychosis, manic-type or manic, depressive
psychosis, circular-type but currently manic according to the International
Classification of Diseases, 9th edition); aged 14–65y | Carbamazepine 100 mg tid Not reported (maximum
dose, 900 mg/d; range 300–900 mg/d) vs. Chloropromazine 50 mg
tid (maximum dose, 450 mg/d; range 150–450 mg/d) for
3–5w | | Psychotropic drugs prohibited; hypnotics allowed
at bedtime | Physician assessment of degree of illness (0-
normal, 4-extremely severe) and improvement (0–6) weekly;
Clinical Psychopharmacology Research Group (CPRG) scale assessed mania
weekly | Age: 35.5 Male
(%):32(53.3%) Ethnicity: Not reported | Monopolar mania:
11(18.3%) Bipolar mania:
41(68.3%) Mixed mania: 1(1.7%) First
mania attack: 7(11.7%)
NSD between groups in
regards to background characteristics (data NR)
Carbamazepine
vs. Chloropromazine Moderate severity: 16(50%) vs.
12(43%) Severe severity: 14(44%) vs.
12(43%) Symtom rating scale: 74.5 vs. 71.6 | Number screened not reported/number eligible not
reported/63 enrolled/63 randomized | 8 withdrawn/lost to follow-up not reported/55
analyzed | Carbamazepine vs. Chloropromazine Marked
improvement: 12(40%) vs. 5(20%) Moderate
marked improvement: 21(70%) vs.
15(60%)
Improvement slight higher in
carbamazepine group though not significant
Rate of
improvement: 1w: 20% vs. 13% 2w:
45% vs. 43% 3w: 65% vs.
52% 4w: 76% vs. 60% 5w:
77% vs. 54% 6w: 71% vs.
56% | Carbamazepine vs. Chloropromazine Onset of
improvement: 3d: 4(12.5%) vs.
2(7.1%) 4–7d: 15(46.9%) vs.
11(39.3%) 8–10d: 2(6.3%) vs.
1(3.6%) 11–14d: 1(3.1%) vs.
2(7.1%) 15d: 10(31.3%) vs.
12(42.9%)
NSD between groups in regard to onset
of improvement | No correlation between overall therapeutic
response and any patient characteristics (e.g., age); nor between overall
therapeutic response and treatment group | Carbamazepine vs. Chloropromazine Symptom
rating scale: 44.3 vs. 47.8 Difference in scores: 30.2 vs.
23.8
Trend toward greater decrease in rating score observed
in carbamazepine group, NSD between groups in overall changes in rating
scale | Side effects evaluated weekly using scale (0-none,
3- severe); laboratory measures taken weekly to determine if drug should be
discontinued in the event of serious AEs | Carbamazepine vs. Chloropromazine AEs:
59% vs. 86%; p<0.05 Drowsiness:
29% vs. 59%; p<0.05 Headache:
26% vs. 31% Cutaneous symptoms:
16% vs. Dry mouth: 15% vs.
24% Lassitude: 15% vs.
21% Dizziness: 12% vs.
28% Orthostatic hypotension: 0% vs.
17% Weakness: 0% vs.
17% Hypersalivation: 0% vs.
17% Nasal stuffiness: 0% vs. 17%;
p<0.05
Abnormal changes in serum glutamic oxaloacetic
transaminase, serum glutamic pyruvic transaminase and leukocyte count in
both groups; none disclosed any serious AE suggestive of kidney, liver or
heart damage | Total withdrawals:
8/63(12.7%) Withdrawals due to AEs:
4/63(9.5%) (due to blurred vision, fever, exanthenia and leg
numbness) | |
Post, 1987 Canada | DB, single-center Inpatient setting | Bipolar Manic by the Research Diagnostic
Criteria and by the DSM-III criteria | Carbamazepine 200–400 mg/d (titrated
against side effects, maximum dosage of 1600–2000 mg/d; mean
dose 1242 mg/d) vs. Neuroleptics (chloropromazine, thioridazine, pimozide)
vs. Lithium for 21d | 2d placebo-controlled phase before administration
of medication | None | Consensus ratings twice daily by nurse-observers
using the Bunney-Hamburg scale for depression, mania, anxiety, anger and
psychosis | Age: Not reported Male (%): Not
reported Ethnicity: Not reported | Responders (n=12) vs. Nonresponders
(n=9) MRS: 8.0(1.4) vs. 3.4(2.2);
p<0.001 Age on onset: 20.3(4.9) vs. 17.9(4.7);
p=0.30 Duration of illness: 18.3(11.4) vs. 15.3(12.1);
p=0.59 Total episodes: 42.5(50.3) vs. 18.2(17.3);
p=0.15 Prior illness (w): 288.8(258.9) vs. 269.1(275.0);
p=0.88 Prior hospitalization (w): 156.5(297.9) vs.
61.7(56.6); p=0.31 Prior manic episodes: 23.5(29.0); vs.
10.3(8.6); p=0.17 Prior depressive episodes: 19.0(21.7)
vs. 7.8(8.9); p=0.25 Episodes/year: 4.4(6.0) vs.
1.2(0.8); p=0.098 Episodes in prior year: 7.0(5.6) vs.
2.7(2.4); p=0.03 Dysphoric: 6.2(1.9) vs. 4.5(2.1);
p<0.10 | Number screened not reported/number eligible not
reported/19 enrolled/19 randomized | 0 withdrawn/0 lost to follow-up/19 analyzed | Degree of improvement with carbamazepine similar
to those treated with neuroleptics and lithium in previous
studies
Response (≥ 2 ponts in mania ratings):
12(63%)
Responders were signficantly more manic
during placebo phase than nonresponders; more dysphroic and more rapid
cyclers than nonresponders; even though there was marked to moderate
response, they were as symptomatic as non-responders
Response
correlated with degree of psychosis (r=0.41, p<0.09) and
anxiety (r=0.46, p<0.05)
No relationship
between carbamazepne blood levels to degree of clinical antimanic
response | 8/9 patients with placebo discontinuation trials
led to relapse of manic or psychotic symptomalogy; a second carbamazepine
trial lead to improvement again
Robust response in 6/7
patients receiving lithium in addition to carbamazepine at end of trial | Responders (n=12) vs. Nonresponders
(n=9) MRS < 2.5: 7(58.3%) vs.
2(22.2%); NSD
Sleep improvement significantly
improvement in responders (p<0.001) than non-responders
(between-group, p<0.01) | Negative history for affective illness in first
degree relatives with responders; positive history equally distributed
between groups (p=0.04) | Not reported | No AEs reported
No patients had EEG
abnormalities | Total withdrawals:
0(0%) Withdrawals due to AEs: 0(0%) | Improvement on carbamazepine was not always
complete within this time frame |
Revicki, 2005 U.S. | RCT, multicenter, open-label, parallel-group
study Inpatient and outpatient setting | Patients ≥ 18y with DSM-IV diagnosis
of bipolar disorder and hospitalized for treatment of acute manic or mixed
episode; female patients of childbearing age must be using effective birth
control | Divalproex sodium 15–20 mg/kg/d
(titrated up to optimize clinical response; average dose 1504 mg/d) vs.
Lithium (1800 mg/d for acute mania, between 900– 1200 mg/d for
maintenance; average dose 1213 mg/d) for 12m
Divalproex vs.
Lithium Average follow-up (d): 306(119) vs. 339(85);
p=0.025 | Not reported | Anticoagulants were prohibited; lithium and
divalproex allowed in combination with other study drug; during initial
hospital stay most patients received an antipsychotic combined with study
medication; some patients continued with antipsychotics or
antidepressants | Mania Rating Scale and the
Depressive Syndrome Scale at baseline and at discharge; QoL and
clinical outcomes assessed every 3m; interviews conducted by investigators
using the World Health Organization Composite Internation Diagnostic
Interview (CIDI); Medical Outcomes Stuy 36-item short-form Health Survey,
the Mental Health Index to report additional QoL and other patient ouc | Divalproex vs. Lithium Age: 33.8(13.8) vs.
37.4(11.7) Male (%):50(48.1%) vs.
48(49.5%) Ethnicity: White: 62(59.6%) vs.
56(57.7%) | Divalproex vs. Lithium Rapid cycling:
11(11.1%) vs. 13(13.4%) Mixed mania:
50(48.1%) vs. 54(55.2%) Alcohol abuse:
46(44.2%) vs. 47(48.2%) Drug abuse:
45(43.3%) vs. 48(49.5%) Hospitalization
during acute phase (d): 11.0(9.4) vs. 11.7(9.6)
MRS:
22.0(8.9) vs. 23.2(8.5) DSS: 24.0(7.3) vs. 25.2(8.0) SF-36,
physical component: 484.3(9.6) vs. 49.4(9.8) SF-36, mental component:
40.4(14.2) vs. 39.3(14.0)
Previous lithium:
68(65%) vs. 61(63%) Previous divalproex:
45(43%) vs. 31(32%) | Number screened not reported/number eligible not
reported/221 enrolled/221 randomized | 142 withdrawn/47 lost to follow-up/172
analyzed
Only 201/221 entered maintenance phase (between-
group, p=0.051 for available follow-up data) -- | Divalproex vs. Lithium Monotherapy: 3m:
26% vs. 22% >3m: 69% vs.
63%
Use of combination with lithium or
divalproex: 3m: 14% vsl 18% >3m:
3% vs. 6%
Use of antidepressants or
antipsychotics: 3m: 47% vs. 50% >3m:
29% vs. 30% | Divalproex vs. Lithium No mania and
depression (m): 5.3(4.6) vs. 5.4(4.4);
p=0.814
Physical component: 1m: 48.2(11.2)
vs. 50.4(9.8) 3m: 49.0(10.7) vs. 50.7(11.1) 6m: 49.8(11.2) vs.
49.9(10.1) 9m: 48.8(11.4) vs. 49.1(10.1) 12m: 49.1(10.9) vs.
50.4(9.8)
Mental component: 1m: 42.0(14.4) vs.
42.2(12.9) 3m: 41.8(13.4) vs. 40.4(14.1) 6m: 43.5(13.8) vs.
42.6(12.5) 9m: 44.7(12.9) vs. 43.2(13.7) 12m: 44.1(12.6) vs.
43.1(12.2)
NSD between groups based on the MCS; or base on
the PCS or measures of disability days (data NR) | Divalproex vs. Lithium Mental
health: 1m: 64.5(19.0) vs. 65.8(17.1) 3m: 66.0(18.1) vs.
63.1(19.5) 6m: 66.3(18.2) vs. 64.5(17.1) 9m: 67.6(17.3) vs.
65.1(19.0) 12m: 66.8(18.8) vs. 65.7(17.1)
Time to
first hospitalization not significant between groups
(p=0.616)
Hospital stay (d): 11.3(21.4) vs.
14.1(27.7); p=0.417
Mean hospital stay was lower
in patients continuing therapy than those who discontinued
(p=0.016) | Use of mood stabilizers (n=129) vs
no-use of mood stablizers (n=72): MCS scores: 6m: 43.7
vs. 40.7; p=0.194 9m: 44.9 vs. 40.7;
p=0.057 12m: 44.0 vs. 41.9;
p=0.280
Restricted activity (d): 12.8(2.4) vs.
23.6(4.8); p=0.048 | AEs recorded during acute phase; during 12m
maintenance phase, AEs monitored | No AEs reported | Divalproex vs. Lithium Total withdrawals:
142/221 Withdrawals due to AEs: 7/104(7%) vs.
14/97(14%) | Medical resource use and costs were collected;
study medication costs higher in divalproex group, but medical and inpatient
costs were low (p=0.693)
NSD between early
dropouts (n=17) and those who continued based on gender, race,
age, hospitalization; only significant in education (p=0.049),
nocompliance (p<0.0001), suicidality (p=0.005) and
discharge setting (p=0.046) |
Greil, 1997 Germany MAP Study
(Multicenter study of long-term treatment of affective and schizoaffective
psychoses study) (Poor) | Multicenter, open-label, long- term
RCT Initially inpatient at psychiatric university hospitals then
outpatient setting | Current episode of bipolar affective or
schizoaffective disorder (ICD-9, World Health Organization, 1978; DSM was
not a diagnostic criterion but patients were assessed with DSM); at least
one former episode during the 3 y (schizoaffective patients) or 4 y (bipolar
patients) preceding the index episode; no preventive treatment immediately
before onset of present episode; age 18 to 65 y; no current alcohol or drug
abuse. Patients in stable condition (Global Assessment Score (GAS) >
70 for at least 2 wk after discharge) entered the maintenance phase. Data
presented for patients with bipolar disorder only. | Carbamazepine - mean dose 635
+/− 190 mg/d (between month 2 and study termination;
dosing schedule not reported) vs. Lithium - mean dose 26.8
+/− 6.76 mmol/l (between month 2 and study
termination; dosing schedule not reported) for 2.5 y | None | Antidepressants, neuroleptics,
benzodiazepines | 6-point psychopathology scale (1 = no
disturbance, 6 = extremely severe recurrence) and 4-point
Morbidity Index (0 = no symptoms, 3 =
hospitalization) at beginning of maintenance phase, 3 times within first 3
months, every 8 to 12 weeks, then at 1, 2, and 2.5 years and between
outpatient appointments as needed.
Main outcomes of interest
were criteria for failure: (a) Hospitalization; (b) Recurrence
(psychopathology scale rating of 5 ("recurrence") or 6 ("extremely severe
recurrence") of an affective episode (RDC criteria); (c) Recurrence and/or
concomitant psychotropic medication (needed for at least 6 mo); (d)
Recurrence and/or concomitant psychotropic medication and/or severe adverse
events (prompting discontinuation) | Carbamazepine vs. Lithium Age, mean (SD),
y: 42 (14) vs. 45 (14) Male/Female: 46%/54%
vs. 50%/50% Ethnicity not reported | Carbamazepine (N = 70) vs. Lithium (N
= 74) 91% of the ICD-9 diagnosed patients fulfilled
the DSM-III-R criteria of a bipolar disorder (58% were pure
"Bipolar," corresponding to Bipolar I (DSM-IV); 33% were
"Bipolar NOS")
Age at onset, mean (SD), y: 32.8 (12.8) vs.
35.4 (13.1) Suicide attempts (% of
patients) None: 66% vs.
57% 1: 23% vs.
30% 2 or more: 11% vs.
13% Episodes of illness
(%) 2: 22% vs.
8% 3-5: 34% vs.
51% 6 or more: 44% vs.
41% Hospitalization
(%) 1–2: 34% vs.
29% 3–6: 57% vs.
62% 7 or more: 8% vs.
10% | Number screened not reported/375 eligible/175
enrolled/144 randomized | 41 withdrew/None lost to follow- up/144
analyzed | Carbamazepine (N = 70) vs. Lithium (N
= 74) (ITT Analysis) Events (number of
failures) Hospitalization: 14 vs. 13 Recurrence:
20 vs. 17 Recurrence and/or concomitant medication : 27 vs. 22 (p
= 0.041) Recurrence and/or concomitant medication and/or
severe adverse events: 36 vs. 26 (p = 0.007) | Kaplan-Meier estimates of survivor functions (ITT
Analysis) were similar for hospitalization and recurrence, and showed a
higher cumulative proportion of patients remaining well on lithium than
carbamazepine for recurrence/concomitant medication and
recurrence/concomitant medication/severe adverse
events.
Similar results were found when DSM-III-R diagnoses
of "Bipolar Disorders" (including "Bipolar Disorder NOS") were used. | Frequencies of treatment failures/per-protocol
completers Hospitalization: 14/40 (35%) vs. 13/60
(22%) (p = 0.17) Recurrence: 20/43
(47%) vs. 17/60 (28%) (p =
0.06) Recurrence/concomitant medication: 27/46 (59%) vs.
22/60 (37%) (p = 0.03) Recurrence/concomitant
medication/severe adverse events: 36/55 (65%) vs. 26/64
(41%) (p = 0.01) | Amount of concomitant medication (antidepressants,
neuroleptics, benzodiazepines), arithmetic means of Defined Daily Doses
(agreed upon standard doses, often close to the manufacturer- recommended
average daily doses) At 1 y: 1.60 vs. 1.27 At 2 y: 1.24 vs.
0.90 At 2.5 y: 1.38 vs. 1.67 (NSD for each
analysis)
About 70% of patients did not receive
additional medication. | Monitored | Carbamazepine vs. Lithium
Adverse
events leading to withdrawal, n Carbamazepine: exanthema
[allergic skin rashes] (6), enlarged lymph nodes
with exanthema (1), diarrhea (1), hepatopathy (1) Lithium: acne and
weight gain (1), psoriasis (1), nausea (1), disturbance of potency
(1) Pattern of withdrawals due to adverse events: 7/9 withdrawals in
carbamazepine group occurred in the first 4 mo vs. 4/4 withdrawals in
lithium group occurred after 3, 4, 5, and 25 mo.
Adverse
events more frequent on lithium Slight tremor (12% vs.
37%; p < 0.002) Polydipsia (6% vs.
32%; p < 0.001) Polyuria (10% vs.
29%; p = 0.009) Diarrhea (10% vs.
28%; p = 0.015)
Adverse event more
frequent on carbamazepine Pruritus (20% vs.
7%; p = 0.046)
Suicides: 1 committed
and 1 attempted suicide (both on carbamazepine) | Carbamazepine vs. Lithium Total
withdrawals: 27/70 (38.6%) vs. 14/74
(18.9%) Withdrawals due to adverse events: 9/70
(12.9%) vs. 4/74 (5.4%) | Open-label design. |
Greil, 1998 Germany, Switzerland MAP
Study (Poor) | Same as Greil, 1997; supplemental evaluation using
DSM-IV terminology and post hoc "classical" and "nonclassical" subgroups
Outpatient setting | Same as Greil, 1997; bipolar I, II or NOS (DSM-IV)
required prophylactic treatment | Same as Greil, 1997 | None | Same as Greil, 1997 | Kaplan-Meier surivivor estimated. Fisher exact
test, Tarone-Wave statistics test. Mantel-Haenszel statistics. Main
outcomes: Hospitalization; recurrence; recurrence and/or concomitant
psychotropic medication (antidepressants and/or neuroleptics) for at least 6
mo; recurrence and/or concomitant psychotropic medication and/or side
effects prompting discontinuation of treatment; and recurrence and/or
subclinical recurrence | Not reported | Not reported | Numbers screened, eligible, and enrolled were not
reported/171 randomized | 40/171 (23.4%) withdrew/None lost to
follow-up/171 (ITT) or 80 (Per-Protocol) analyzed (see Kleindienst,
2002) | Classical bipolar subgroup (ITT
analysis) Carbamazepine (N = 32) vs. Lithium
(N = 35) Hospitalizations: Lithium was superior to
carbamazepine using Kaplan-Meier survival estimates (p = 0.005);
cumulative survival at 30 mo (estimated from figure): 50% vs.
78% Lithium superior to carbamazepine for other failure
criteria (data not reported) Recurrence: p =
0.010 Recurrence/concomitant medication: p =
0.002 Recurrence/concomitant medication/severe adverse events: p
< 0.001 Recurrence/subclinical recurrence: p <
0.001 | Nonclassical bipolar
subgroup Carbamazepine (N = 53) vs. Lithium (N
= 51) Hospitalizations: NSD using Kaplan-Meier survival
estimates (p = 0.075); cumulative survival at 30 mo (estimated
from figure): 70% vs. 60% NSD was found for
the other failure criteria | Carbamazepine and Lithium Risk for
treatment failure compared with a classical bipolar patient with one (at
least 2) nonclassical diagnostic feature(s) Hospitalization: 0.54
(0.40) (p < 0.05) and 1.42 (2.52) (p < 0.05) Recurrence: 0.75
(0.40) (p < 0.1) and 1.34 (2.20) (p <
0.1) Recurrence/concomitant medication: 0.88 (0.53) and 1.42 (1.89)
(p < 0.1) Recurrence/concomitant medication/severe adverse
events: 0.91 (0.51) and 1.50 (1.98) (p <
0.05) Recurrence/subclinical recurrence: 0.76 (0.82) and 1.35 (2.43)
(p < 0.05) | | Not reported | Not reported | Total withdrawals: 28/85 (32.9%) vs.
12/86 (14.0%) (before suffering recurrence; p =
0.004) | There were numerous threats to internal validity:
classification of patients into classical and nonclassical bipolar subgroups
was done post hoc; nonclassical subgroup analysis may have been
underpowered; no statistical adjustment for multiple comparisons; open-label
design. |
Greil, 1999 “bipolar
II/NOS” Germany MAP Study (Poor) | Same as Greil, 1997 | Same as Greil, 1997, except that this report
describes patients with bipolar II disorder or bipolar disorder NOS
according to DSM-IV (these patients were originally classified as bipolar
disorder NOS under DSM-III-R) | Same as Greil, 1997 | None | Same as Greil, 1997 | Global psychopathology rating scale (1
= no disturbance, 4 = subclinical recurrence, 5
= recurrence, or 6 = extremely severe
recurrence). Main outcomes of interest were criteria for failure: (a)
Hospitalization; (b) Recurrence (psychopathology scale rating of 5 or 6 of
an affective episode (RDC criteria); (c) Recurrence and/or concomitant
psychotropic medication for at least 6 mo; (d) Recurrence and/or concomitant
psychotropic medication and/or adverse events prompting discontinuation; and
(e) recurrence and/or subclinical recurrence (score of 4, 5, or 6). Surval
Analysis (Kaplan-Meier estimates of the survivor functions) 2.5 years
period. | Age, mean, y: 41 Female: 60% Ethnicity
not reported | Not reported | Not reported/Not reported/Not reported/57 (This
study describes patients with bipolar II disorder or bipolar disorder not
otherwise specified (NOS) (DSM-IV), who were previously classified as
bipolar disorder NOS under DSM- III-R). Thus, this is a subgroup of the
population described in Greil, 1997 | 18 withdrew/Number lost to follow-up not
reported/57 analyzed in ITT survival analyses; number not reported for per-
protocol completer analysis | CCarbamazepine vs.
Lithium
Frequency of failures/completers for
failure criteria, relative risk (RR) Hospitalization:
3/18 (17%) vs. 7/21 (33%), RR = 0.50 (p
= 0.29) Recurrence: 5/18 (28%) vs. 8/21
(38%), RR = 0.73 (p =
0.73) Recurrence and/or concomitant medication: 10/19
(53%) vs. 10/21 (48%), RR = 1.11 (p
= 1.00) Recurrence and/or concomitant medication and/or
severe adverse events: 12/21 (57%) vs. 12/22 (52%),
RR = 0.91 (p = 1.00) Recurrence and/or
subclinical recurrence: 11/20 (55%) vs. 17/24 (71%),
RR = 0.78 (0 = 0.35) Survival time was
significantly higher under lithium than under carbamazepine
(p=0.03) | NSD in survival times by Kaplan- Meier estimates
(ITT, p = 0.17 to 0.94) | | | Not reported | Not reported | Carbamazepine vs. Lithium Total withdrawals: 11/29
(38%) vs. 7/28 (25%) Withdrawals due to
adverse events: Not reported | Open-label design. It is not clear whether the
subgroup analysis was decided a priori or post hoc. Adjustment for
multiple testing was not reported. Because of the naturalistic (open-
label) study design, generalizability may be possible. |
Greil, 1999 (-- “bipolar
I”) Germany MAP Study (Poor) | Same as Greil, 1997 | Same as Greil, 1997; also bipolar I disorder
(DSM-IV, corresponding to bipolar disorder under DSM-III-R) | Same as Greil, 1997 | None | Same as Greil, 1997 | Psychopathology severity and type rating scale (1
= no disturbance, 4 = subclinical recurrence, 5
= recurrence, 6 = extremely severe recurrence)
monthly. Criteria for treatment failure: (a) hospitalization; (b)
recurrence (psychopathology rating of 5 or 6); (c) recurrence and/or
concomitant psychotropic medication for at least 6 mo; (d) recurrence and/or
concomitant psychotropic medication and/or side effects prompting
discontinuation of treatment; and (e) recurrence and/or subclinical
recurrence (psychopathology rating of 4, 5, or 6) | Age, mean, y: 40 Male/Female:
50%/50% Ethnicity not reported | 171 patients met DSM-IV diagnosis of bipolar
disorder; 114 had bipolar I disorder | 171/114/114/114 | 22 withdrew/Number lost to follow-up not
reported/114 analyzed in Kaplan-Meier survival analyses; up to 103
completers analyzed for failure rates | Carbamazepine vs.
Lithium
Failure rates, relative risk
(RR) Hospitalization: 21/38 (55%) vs. 20/54
(37%), RR 1.49 (p = 0.09) Recurrence: 23/39
(59%) vs. 21/53 (40%), RR 1.49 (p =
0.09) Recurrence/concomitant medication: 28/42 (67%) vs.
24/54 (44%), RR 1.52 (p = 0.04)
[calculated NNt (95% CI): 5
(2.36) Recurrence/concomitant medication/severe adverse events: 34/48
(71%) vs. 25/55 (46%), RR 1.54 (p =
0.01) [ calculated NNt (95% CI): 4
(2.14)] Recurrence/subclinical recurrence: 31/44
(71%) vs. 29/56 (48%), RR 1.48 (p =
0.04) Note: There appears to be an error: 29156 does not equal
48%, but equals 52% this produces a nonsignificant
RR of 1.46 (p = 0.06) | Symptomatology leading to rehospitalization
Depression/mania/other: 37%/21%/42% vs.
38%/31%/31% (NSD) Kaplan-Meyer
survival for clinical or subclinical recurrence at 30 mo, estimated 0.34 vs.
0.55 (p = 0.03) | | | Not reported | Not reported | Total withdrawals: 17/56 (30%) vs.
5/58 (8%) Withdrawals due to adverse events: Not
reported | Open-label design. It is not clear whether the
subgroup analysis was decided a priori or post hoc. Adjustment for multiple
testing was not reported. |
Kleindienst, 2002 Germany,
Switzerland MAP Study (Poor) | Same as Greil, 1997; supplemental evaluation of
inter-episodic morbidity and dropout Outpatient setting | Same as Greil, 1997. Patients with bipolar
affective disorder (DSM-IV) were analyzed in this report. | Same as Greil, 1997 | None | Same as Greil, 1997 | Morbidity Index (MI) (for assessing recurrences
leading to re-hospitalization and inter-episodic symptoms); retrospective
symptomatology scale (manic, depressive, mixed, schizoaffective, or other);
4-point severity scale (0 = no affective symptoms; 3
= affective symptoms that necessitate hospitalization);
dropouts; KK-Scale for illness concepts; Munich Personality Test for pre-
morbid personality every 8 to 12 wk
Good responders
= average inter-episodic morbidity below the median, no re-
hospitalization, no dropout | Carbamazepine (N = 85) vs. Lithium (N
= 86) Age, mean (SD), y: 39 (13) vs. 41 (13) Male
/ Female: 42% / 58% vs. 45% /
55% Ethnicity not reported | Number of previous episodes, mean (SD): 3.27
(2.32) vs. 3.07 (2.22) GAS score, mean (SD): 79 (10) vs. 79
(10) Psychiatric comorbidity: 16% vs.
16% Pre-morbid personality scores were similar between
treatment groups except for Extraversion, mean (SD): 13.5 (5.7) vs. 11.2
(6.6); p < 0.05 | Numbers screened, eligible, and enrolled were not
reported / 171 randomized | 40/171 (23.4%) withdrew / None lost to
follow-up / 171 (ITT) or 80 (Per-Protocol) analyzed | Carbamazepine vs. Lithium
Dropouts:
29/85 (34.1%) vs. 11/86 (12.8%) (p =
0.001) Dropouts mostly related to treatment,n: 26 vs.
10
Re-hospitalization: 28% vs. 31%
(p=0.74) | % of time between affective episodes:
42% vs. 36% Inter-episodic symptomatology requiring
treatment; 64% vs. 60%
Average
inter-episodic morbidity correlated with re-hospitalization: r =
0.22 (p = 0.045) vs. r = 0.34 (p =
0.0013)
Average inter-episodic morbidity index over time,
first vs. last 6 mo Carbamazepine: 0.54 vs. 0.44 (p =
0.11) Lithium: 0.54 vs. 0.30 (p = 0 0051) | Good responders (ITT): 20/85 (23.5%)
vs. 34/86 (39.5%) (p = 0.032). | | Not reported | Not reported | Total withdrawals: 29/85 (34.1%) vs.
11/86 (12.8%) Withdrawals due to adverse events: 8/85
(9.4%) vs. 3/86 (3.5%) | The study took place when carbamazepine was
relatively new to mood disorders; therefore, open-label design may have
biased against carbamazepine because of unfamiliarity with the
drug. The principal goals and contribution of this study were the
refined evaluations of drop- outs and of subthreshold symptomatology.
However, it is unclear whether these analyses were planned a priori. |
Hartong, 2003 The
Netherlands (Fair) | Multicenter Double-blind, double-dummy RCT 18
outpatient clinics | Bipolar disorder (DSM-III-R criteria) with at
least 2 symptomatic episodes during the previous 3 yr; no antidepressants,
antipsychotics, or benzodiazepines above allowed dosages; at least 18 yr
old; Dutch- speaking. Report excluded 6 schizoaffective patients who
had been recruited per protocol. Total of less than 6 months of
previous lithium or carbamazepine treatment | Lithium 400 to 800 mg/d, then titrated to blood
concentrations between 0.6 and 1.0 mmol/l vs. Carbamazepine 200 to 400 mg/d,
then titrated to blood concentrations between 6 and 10 mg/l for 2 yr | Run-in acutely randomized patients on double-blind
treatment; entered actual prophylactile phase after recovery from acute
episode. | Benzodiazepines at doses equivalent to a maximum
of 50 mg/d of oxazepam. For impending relapse, doses equivalent to a
maximum of 100 mg/d of oxazepam were allowed for up to 14
days. Medications for somatic diseases (not specified). | Recurrence of an episode of (hypo)mania or major
depression (DSM-III-R criteria) (Primary Outcome Measure); Comprehensive
Psychiatric Rating Scale (CPRS); Bech Rafaelsen mania Scale (BRMAS), Bech
Rafaelsen M,elancholia Scale (BRMES) at baseline then every month. | Mean age (SD) 41.9 (13.9) 45.7%
male, 54.3% female Ethnicity not reported | Bipolar I 72/94 Bipolar II
22/94 Rapid Cycling 10/94 Non-rapid cycling 84/94 | --/--/150/144 | 46 withdrawn/0 lost to follow-up/94 analyzed | Lithium vs
Carbamazepine
Recurrence: 27.3% vs.
42.0% (p- value not reported)
Episodes on lithium
primarily occurred in first 3 months (hazard 0.3 at 100 d) while risk with
carbamazepine was 40%/yr. Dropped out: 36.4%
vs. 26.0% Completed 2 yr without episode:
36.4% vs. 32.0% (p-value not reported) | Recurrence, prophylactically randomized patients:
14.3% vs. 46.7%. Recurrence, acutely
randomized patients: 42.8% vs. 35.0%. About
40% of these patients experienced an episode within the first 3
mo on lithium. Thereafter, the risk of recurrence with lithium was
< 10%/y. | Recurrence in prophylactically randomized patients
with (hypo)manic index episode: 0% vs. 61.5% (p
< 0.01) Recurrence in bipolar II patients: 0% vs.
50.0% (p < 0.05) | | Monitored | Lithium vs. Carbamazepine AEs with >
10% treatment difference at 2 wk (N = 88): Blurred
vision 26% vs. 11% Difficulty concentrating
45% vs. 33% Feeling thirsty 41%
vs. 22% Decreased appetite 21% vs.
9% Hand tremor 31% vs.
4% Muscular weakness 14% vs.
4% Increased appetite 17% vs.
33% | Lithium vs. Carbamazepine: Total withdrawals:
16/44 (36.4%) VS. 13/50 (26.0%) Withdrawals
due to adverse events: 5/144 (3.5%) vs. 4/144
(8%) | Two randomization points: prophylactically
randomized (at start of prophylactic treatment phase, the actual study
entry) or acutely randomized (during an acute episode of (hypo)mania or
depression). Uneven randomization with more patients prophylactically
randomized to carbamazepine (n = 30) than lithium (n
= 23). Few bipolar II patients were acutely randomized
and they were unequally distributed between treatments. Did not
incorporate secondary outcome measures a priori. The proportional hazard
assumption did not hold; therefore, instead of the intended Kaplan-Meier
analysis, post hoc sensitivity analyses were performed. |
Lerer, 1987 U.S. (Poor) | Double-blind, double-dummy, parallel-group RCT
Outpatient and inpatient setting | Bipolar disorder, manic (DSM-III); age 21 to 65 y;
physically healthy without seizure disorder | Carbamazepine starting at 600 mg/d and titrated to
serum concentration of 8 to 12 μg/ml vs. Lithium starting at 900
mg/d and titrated to serum concentration of 1.0 mEq/l for 4 wk | 7- to 14-d washout of psychotropic medications
other than chloral hydrate or barbiturates for sedation | Chloral hydrate or barbiturates for sedation | Clinical Global Impression (CGI) scale; Brief
Psychiatric Rating Scale (BPRS); Beigel- Murphy Manic State Rating SCale
(MSRS) at baseline and weekly thereafter | Carbamazepine (N = 14) vs. Lithium (N
= 14) (Completer Population) Age, median, y: 44 vs.
37 Male / Female: 57.1% / 42.9% vs.
35.7% / 64.3% Ethnicity not reported | Previous response to lithium: Moderate/Good 6
(42.9%) vs. 9 (64.3%) | Number screened and eligible not reported / 34
enrolled / 34 randomized | 6 withdrew / None lost to follow-up / 28
analyzed | Carbamazepine vs. Lithium
Change in
mean BPRS score, baseline to wk 4 (estimated from figure): -6 vs.
-10 Calculated difference between changes in mean scores: 4 (NSD for
improvement scores, data not reported) Individual BPRS items with
significant treatment differences: --hostility (p <
0.05) --hostility-suspiciousness factor (p < 0 01) | Change in mean MSRS, baseline to wk 4 (estimated
from figure): - 50 vs. -101 Calculated difference between changes in
mean scores: 51 (NSD for improvement in MSRS scores, data not reported) | Mean CGI change in severity of illness scores,
baseline minus wk 4 (estimated from figure): 1.3 vs. 2.6 (p <
0.05) | | Monitoring | Carbamazepine (n): reversible increase in liver
enzyme test results > 4 to 6 times above normal (1); hepatitis,
consistent with drug-induced type (1); severe pruritic maculopapular rash
(1) decreased white blood cell count (1). Overall, there was a mean (SD)
decreased in WBC count of 35% (from baseline of 8143 (3438.7) ml
to 5264 (1801) ml.
Lithium (n): tremor and nausea (1);
pruritic maculopapular rash (1); drowsiness and slured speech (2) | Unable to determine because of discrepancies in
data | Cannot exclude the possibility of a type II error. |
Lusznat, 1988 U.K. (Poor) | Double-blind, double-dummy, parallel-group RCTwith
6-wk acute trial then 12-month follow-up Initially inpatient then
outpatient setting affiliated with a Dept. of Psychiatry | Confirmed diagnosis of mania or hypomania; age 17
to 64 y; Bech- Rafaelson mania rating scale score >/=
10 | Carbamazepine (starting at 200 mg/d and titrating
to serum concentration of 0.6 to 1.2 mg/dl) vs. Lithium (starting at 400
mg/d and titrating to serum concentration of 0.6 to 1.4 mmol/l) for 18
mo | None | Neuroleptics had been given to 52 patients prior
to baseline assessment and during acute trial. Hypnotics (usually
temazepam), antidepressants, or neuroleptics during follow-up trial. | Bech-Rafaelsen Mania Rating Scale (B-R MRS), side
effect rating scale (ranging from 0 to 2, 13 or more symptoms); 16-h
Dexamethasone Suppression Test (DST) at baseline, 3–4 d after
starting medication, then at 1 wk and weekly until week 6. Global rating of
severity of mania, B-R MRS, side effecting rating, Hamilton Rating Scale for
Depression (HRSD, 17 items) when global rating of mania was 0, and rescue
medications monthly for a year. | Not reported | DSM-III diagnosis, n: Schizoaffective (2), bipolar
without psychotic features (35)
Carbamazepine vs.
Lithium History of alcohol abuse, n: 8 vs. 4 B-R MRS score:
15.8 vs. 14.6 | 128 screened/54 eligible/54 enrolled/54
randomized | 25 withdrawn/Lost to follow-up none/Number
analyzed for B-R MRS scores not reported | Carbamazepine vs. Lithium
B-R MRS
score, calculated change in mean B-R MRS score from baseline to wk 6,
estimated: −12 vs. − 13 (NSD) HRSD scores:
NSD (data not reported) Daily neuroleptic dose, calculated change in
mean daily neuroleptic dose from baseline to wk 6, estimated, mg/d:
−700 vs. −800 (NSD) | Length of hospital stay, mean (SD), d: 30 (22) vs.
32 (28) (NSD) | Follow-up trial: B-R MRS score, time point not
reported, mean: 1.1 vs. 1.2 (NSD) HRSD scores, mean: 2.9 vs. 3.2
(NSD) Response Predictors to carbamazepine: lower DST at admission (p
< 0.05) Overall result (definitions not
reported) "Poor": 7/27 (25.9%) vs. 12/27
(44.4%) "Satisfactory": 9/27 (33 3%) vs 5/27
(18 5%). | | Monitored and graded on a side effect rating scale
(13 symptoms, rated 0 to 2 according to severity) The mean side
effect rating score was the average of total scores for all
assessments. | Carbamazepine vs. Lithium
Acute
trial Side effect rating scale score, mean: 2.8 vs. 2.8 More
likely reported side effect: Ataxia on carbamazepine vs. Nausea and tremor
on lithium
Follow-up trial Side effect rating scale
score, mean: 1.2 vs. 1.7 NSD) Specific side effects not reported | Only partial data on withdrawals were reported by
treatment Carbamazepine vs. Lithium Total withdrawals: 11/27
(40.7%) vs. 10/27 (37.1%) Withdrawals due to
adverse events: 1/27 (3.7%) vs. 2/27
(7.4%)
Adverse events resulting in
withdrawals Carbamazepine: skin rash Lithium: Seizure, psoriasis
worsened | High rate of drop-outs, which appeared to occur at
random. |
Coxhead, 1992 U.K. (Fair) | Double-blind, double-dummy, placebo-controlled,
parallel- group RCT Outpatient | Current lithium prophylaxis; bipolar disorder
(DSM-III); no other psychotropic medication. | Carbamazepine (starting at 400 mg/d and titrated
to serum concentration of 38 to 51 mmol/l) vs. Lithium (starting at 800 mg/d
and titrated to serum concentration of 0.6 to 1.0 mmol/l) for 1 y | Run-in on previous lithium dose. Patients were
randomized to treatment if, after 4 wk of lithium at previous doses, their
mania rating score remained zero, Hamilton Rating Scale for Depression
(HRSD) score stayed below 4 at −4, −2, and 0 wk, and
no other psychotropic medication was taken. | Temazepam up to 20 mg at night for sedation | Bech-Rafaelsen Mania Rating SCale (B-R MRS), HRSD,
global rating of affective state; rating of duration and severity of mood
changes since previous assessment, recorded at baseline, wk 2, wk 4, then
every 4 wk for 1 y. Affective morbidity index was calculated using
the global ratings of duration and severity of mood changes since previous
assessment. | Carbamazepie (N = 15) vs. Lithium (N
= 16) Age, mean (SD), y: 47 (14) vs. 49
(10) Male/Female: 5/10 vs. 5/11 Ethnicity not reported | Number of previous admissions, mean (SD): 6.1
(3.7) vs. 7.1 (4.6) Duration of illness, mean (SD), y: 17 (11) vs. 17
(14) Nature of last inpatient episode, mania/depression: 11/4 vs.
13/3 | 145 screened/Number eligible not reported/32
enrolled/31 randomized | 2 withdrew/None lost to follow-up/31 analyzed | Carbamazepine (N = 15) vs. Lithium (N
= 16) Relapsed (admitted): 6 (5) vs. 8
(5) Completed (remaining relapse-free at 1 y): 7/15
(46.7%) vs. 7/16 (43.8%) Number of patients
surviving at 3 mo and 1 y: 8 vs. 10 and 7 vs. 7; NSD | Maximum mania and depression scores during
the year (no statistical analyses) B-R MRS, n --0
to 3 (no or few symptoms): 10 vs. 9 --4 to 7 (moderate symptoms):
vs. 1--8 or higher (severe symptoms): 4 vs. 6 | HRSD, n --0 to 5 (mild symptoms): 12 vs.
12 --6 to 11 (moderate symptoms): 3 vs. 2 --12 or higher
(severe 1 symptoms): 1 vs. 1 | Affective morbidity index, mean --Relapsing
(N = 6 vs. 8): 0.86 vs. 0.41 --Completing (N
= 7 vs. 7): 0.12 vs. 0.22 (NSD) | Monitored | Most frequent adverse events Carbamazepine:
drowsiness, dizziness, giddiness, nausea, indigestion (12/15 patients had at
least 1 of these adverse events during the first 4 wk) Lithium:
thirst and/or polyuria (9/16 patients, 56.2%, including 3 severe
cases); weight gain (mean, 4 kg) (9/16 patients, 56.2%) | Total withdrawals: 1/16 (6.2%) vs.
2/15 (13.3%) Withdrawals due to adverse events: 0/16
(0%) vs. 2/15 (13.3%) 2/15 (13.3%) vs.
0/16 (0%) | Primary efficacy variable was not reported.
Negative results may be due to a type II error (small sample
population). |
Small, 1991 U.S. (Poor) | Double-blind, double-dummy, parallel-group RCT
with 2-y double-blind follow-up Tertiary Care Facility; initially inpatient
then 87% discharged to community | Newly hospitalized with bipolar disorder
presenting in manic or mixed phases (diagnosis by Schedule for Affective
Disorders and Schizophrenia-Lifetime version); manic episode (DSM-III- R)
with or without coexisting symptoms of depression; history of at least one
affective episode within the previous 2.5 y; bipolar I disorder (Research
Diagnostic Criteria); score of 7 or more on the manic subsection of the
Depresion and Mania Scale (SDMS-D&M: score range, 3 to 15) and
scores of 60 or less on the Global Assessment Scale (GAS: score range, 1 to
100) | Carbamazepine starting at 200–400
mg/d, titrated until serum concentrations 25– 50 micromol/l vs.
Lithium starting at 300–600 mg/d, titrated until serum
concentration 0.6–1.5 mmol/l for 8 wk. Patients who were
improved or in remission continued to receive double-blind medications for
up to 2 y. | Run-in off therapy following washout of previous
medications and baseline measurements; patients who continued to display
significant psychopathology (Manic Subsection of the Depression and Mania
Scale, SDMS-M, score >/= 7, Global Assessment Scale, GAS,
score </= 60) were randomized. 2-wk washout of previous
lithium and carbamazepine, 1-wk washout of previous neuroleptics | Chloral hydrate for insomnia Amobarbital
for disturbed behavior | SDMS-D&M, GAS, Manic Rating Scale (MRS) of
Young et al., 24-item Hamilton Depression Rating Scale (HDRS), Brief
Psychiatric Rating Scale (BPRS) expanded to include an additional rating of
elevated mood, and Clinical Global Impression Scale (CGIS), recorded at
baseline and weekly; Shopsin- Gershon Social Behavior Checklist, daily for 5
d/wk | Carbamazepine vs. Lithium Age, mean, y:
34.3 vs. 42.6 Male/Female: 41.7%/58.3% vs.
45.8%/54.2% Ethnicity: Not reported | Mean age at onset, y: 23.3 vs. 26.0 No. of
previous episodes of mania,
1–4/5–9/>= 10: 12/10/2 vs.
11/11/2 No. of previous episodes of depression,
1–4/5– 9/>=10: 17/6/1 vs.
14/7/3 Ratio, manic:depressed: 1.4:1 vs. 1.2:1 Lithium treatment of
index episode before admission to study, adequate/inadequate/none, n: 9/12/3
vs. 8/10/6 Scores on Schedule for Affective Disorders and
Schizophrenia-Lifetime version Best level of social relations in past
5 y: 3.0 vs. 3.3 Healthiest overall functioning in past 5 y: 2.9 vs.
2.3 Outcome of last episode: 2.14 vs. 1.92 Comorbid
personality disorders, physical and neurologic problems, and/or hisory of
significant substance abuse, n: 7 vs. 12 | 94 screened/52
eligible/52 Enrolled/52 Randomized | 24 withdrawn at the end of 8 wk (before entering
2- y double-blind phase)/lost to follow-up none/28 analyzed at 8
wk Of 16 who entered long-term phase, 15 withdrew within 2 y/Number
lost to follow-up not reported | Lithium vs. Carbamazepine %
difference in scores MRS: 4% SDMS-M:
-1% SDMS-D: -18% HAM-D:
10 BPRS: 2 CGI-1: 1 GAS: 3 BCL: 8 NSD for
any scores. | Use of as-needed medications at 8 wk, chloral
hydrate/amobarbital, n: 4/17 (23.5%)/4/17 (23.5%)
vs. 3/11 (27.3%)/1/11 (9.1%) | Statistically significant (p < 0.05)
predictors of response to therapy Lithium: None Carbamazepine:
Scores reflecting less psychopathology at baseline: higher GAS score and
lower scores on MRS, BPRS total, CGI- item 1, BPRS Hostile- Suspicious,
SDMS-Manic subsection, and BPRS Thinking-Disturbance | Recurrence during long- term phase, n
(%): 5/8 (62.5%) vs. 3/8 (37.5%)
(statistics not reported) | Monitored with the general inquiry part of the
Systematic Assessment of the treatment of Emergent Events (SAFTEE) | Adverse events leading to withdrawal 2 reported
for Carbamazepine (n): Rash (1) during 8-wk phase, Low granulocyte count (1)
during 2-y double-blind follow-up | Carbamazepine vs. Lithium At wk 8 Total
withdrawals: 7/24 (29.2%) vs. 13/24 (54.2%)
Withdrawals due to adverse events: 0/24 (0%) vs. 1/24
(4.2%)
After wk 8 Total withdrawals: 24/24
(100%) by 24 wk vs. 23/24 (95.8%) by 1 y (NSD)
Withdrawals due to adverse events: 1/8 (12.5%) vs. 0/8
(0.0%)
Withdrawals due to noncompliance during
long- term phase: 2/8 (25.0%) vs. 4/8 (50.0%) | Maintenance of treatment blinding during long-term
phase was tested by asking physicians and nurses to guess the assigned
treatment; accuracy did not reach statistical
significance.
High dropout rates during run-in limits
external validity of study; high dropout rate during long- term follow-up
limited the amount and value of follow-up data. |
Denicoff, 1997 U.S. (Poor) | Double-blind, crossover RCT following open-label
admission phase (average 149.6 +/− 104.1 d)
Outpatient clinics of the National Institute of Mental Health (NIMH),
Bethesda, MD | Bipolar disorder (DSM-III-R) | Phase I or II: Carbamazepine titrated up to 1600
mg/d (target serum concentration: 4 to 12 mg/l) Phase I or II:
Lithium titrated to clinical response (target serum concentration: 0.5 to
1.2 mmol/l) Phase III: Combination Carbamazepine +
Lithium for 1 y per treatment phase (total 3 y of treatment) | Washout - previous carbamazepine or lithium was
tapered over 1 mo if patient had been randomized to the other treatment | Not reported | NIMH-Life Chart Method and Manual prospective
(LCM-p) daily life charting, which included daily mood scale (manic,
depressed, or euthymic) and functional incapacity scale (none, mild,
moderate, or severe), recorded twice daily; average severity score
(calculated by multiplying the number of days at each severity level
[2.5 for mild, 5.0 for moderate, and 10.0 for
severe] and dividing by the number of days in the treatment
phase). Beck Depression Inventory (BDI), Modified Spielberger State-Trait
Anxiety Inventory (MSSTAI), Hamilton Rating Scale for Depression (HAM-D),
Young Mania Rating Scale (YMRS), and Raskin Severity of Depression and Mania
(RSDM) scale, recorded monthly. Clinical Global Impression (CGI) scale,
recorded during treatment phase in comparison with clinical response in the
year prior to the patient taking a mood stabilizer or in the worst year when
patient took ineffective medications.
Relapse was defined as
patient required hospitalization or became
severely incapacitated for at least several days | Age, mean (SD), y: 41.3 (11.4) Male /
Female: 25 / 27 Ethnicity not reported | Employment status: 29 (55.8%) were
employed full-time; 8 (15.4%) were employed part-time; 3
(5.8%) were housewives; 3 (5.8%) were students; 5
(9.6%) were retired; and 4 (7.7%) were not
working. Bipolar II disorder (Research Diagnostic Criteria
[RDC]): 19 (36.5%) Bipolar I
disorder (RDC): 33 (63.5%) (with stipulation that there must be
a full-blown manic episode that led to a hospitalization ro it
sequivalent) History of hospitalization: 39
(75.0%) History of rapid cycling (4 or more episodes in
any 1-year period prior to entering study): 31/51 (60.8%; 1
patient not assessable) History of psychosis: 27
(51.9%) Previous moderate or marked response to Lithium:
16/47 (34%) Carbamazepine monotherapy: 1/4
(25%) Carbamazepine + Lithium: 1/6
(16.7%) | Numbers screened not reported/eligible not
reported/ 52 enrolled / 50 randomized | 21/127 patient episodes of withdrawal (excluding
early discontinuation due to treatment failure) / 6 patient episodes of
dropping out or moved during treatment / 106 patient episodes
analyzed
Note: Since patients crossed over to other
treatments, they were counted as patient episodes in this review. | Carbamazepine vs. Lithium vs.
Combination
CGI marked or moderate improvement (good
treatment response): 31.4% vs. 33.3% vs.
55.2% (NSD)
Percentage of time ill (N
= 29), mean (SD) Mania: 19.0 (19.5) vs. 9.1 (6.8) vs. 8.4
(10.6) (p < 0.01) Depression: 26.3 (22.8) vs. 30.6 (25.3) vs.
29.1 (27.5) (NSD) | Average severity of illness (N = 29),
mean Mania: 0.63 vs. 0.26 vs. 0.25 (p = 0.004; post hoc
analyses showed differences between lithium or combination and
carbamazepine) Depression: 0.93 vs. 1.15 vs. 1.05 (NSD) Total:
1.57 vs. 1.41 vs. 1.30 (NSD)
Number of episodes/year, mean
Mania: 4.55 vs. 3.66 vs. 2.90 (p = 0.041; post hoc analyses
showed differences between combination and either carbamazepine or
lithium) Depression: 2.16 vs. 2.59 vs. 1.74 (NSD) Total: 6.71
vs. 6.25 vs. 4.64 (NSD) | Depression rating scales (score range),
mean HAM-D (0 to 64): 7.8 vs. 7.1 vs. 7.1 (NSD) RSDM
(depression) (3 to 15): 4.9 vs. 4.7 vs. 5.0 (NSD)
BDI (0 to
63): 7.2 vs. 6.9 vs. 7.2 (NSD)
Mania rating scales (score
range), mean YMRS (0 to 60): 5.2 vs. 3.3 vs. 4.4 (NSD) RSDM
(mania) (3 to 15): 4.3 vs. 3.8 vs. 3.9 (NSD) | Correlates of response Predictors
of a... --Positive response to lithium: younger age at study entry;
first treatment by age 20 or earlier; fewer years elapse since onset of
first bipolar symptoms; </= 1 lifetime hospitalization
for mania --Poor response to carbamazepine: > 10 y elapse
between onset of first bipolar symptoms and entry into study and past
history of rapid cycling --Positive response to combination: rapid
cycling; prior course of illness variable reflecting less severity of
illness --Poor response to combination: greater number of
hospitalizations for mania; > 1 hospitalization for mania; greater
mean number of weeks hospitalized per year | Not reported | Adverse events leading to withdrawal
Carbamazepine: rash (9), decreased white blood cell and platelet counts vs.
(1) Lithium (n): cystic acne (1), psoriasis (1) Combination:
None (because patients were not re-exposed to drug if they were
intolerant) | Carbamazepine vs. Lithium Combination, n/N
(%)(where N = no. of patients entering treatment
phase) Total withdrawals: 11/46 (23.9%) vs. 8/50
(16.0%) vs. 2/31 (6.5%) Withdrawals due to
adverse events: 10/46 (21.7%) vs. 2/50 (4.0%) vs.
0/31 (0.0%) | Randomization order was changed in 1 patient.
Research nurses were not necessarily blinded to the third (combination)
phase Selective population of patients previously treated with
carbamazepine or lithium; about 45% of the patients had had
minimal or no response to lithium. |
Bowden, 2000 Canada,
U.S. (Fair) | Multicenter, long-term, double blind,
placebo-controlled, parallel-group RCT with < / = 3-mo
initial open phase followed by 52-wk double- blind randomized maintenance
phase Outpatient setting | Open-label phase: age 18 to 75 yr; bipolar
disorder (DSM-III-R); index manic episode < / = 3 mo
before randomization; at least 1 other manic episode in previous 3
yr
Double-blind phase: scores of < / = 11
on Mania Rating Scale (MRS), < / = 13 on Depressive
Syndrome Scale (DSS), > 60 on Global Assessment Scale (GAS) on 2
consecutive occasions at least 6 d apart. | Open-label stabilization phase:
Investigator’s choice of medication (including divalproex,
lithium, both, or neither) for two consecutive visits at up to 90
d
Double-blind phase: Divalproex (titrated to serum valproate
concentration of 71 to 125 mg/l) vs. Lithium (titrated to serum
concentration of 0.8 to 1.2 mEq/l) for 52 wk | Up to 90-day run-in on investigator’s
choice of medication; patients were randomized if they had, on least 6 d
apart, a Global Assessment Scale (GAS) score > 60, Mania Rating Scale
(MRS) score (derived from the Schedule for Affective Disorders-Change
Version [SADS-C]) < /= 11; and a
Depressive Syndrome Scale (DSS) score (derived from SADS-C) <
14
Washout of psychotropic medication other than lithium or
divalproex before randomization; washout of open-label divalproex and
lithium occurred while blinded drugs were titrated up during first two weeks
of maintenance phase | Lorazepam up to 6 mg/d for 14 d during first month
and no more than 7 d for remainder of study. Haloperidol up to 10 mg/d
during second consecutive wk of lorazepam in first month only. | Time to either a manic or depressive episode
(“any mood episode”) (Primary Outcome Measure); time
to a manic episode; time to a depressive episode; scores on MRS, DSS, and
GAS during maintenance therapy | Divalproex vs. Lithium
vs. Placebo Mean (SD) age, y: 38.9 (12.7) vs. 40.3 (9.8) vs.
38.7 (11.9) 48.8% Male, 51.2% Female
91.3% White, 4.1% Black, 4.6% Other | Divalproex vs. Lithium vs. Placebo MRS,
mean (SD): 3.4 (3.7) vs. 3.2 (3.7) vs. 3.4 (3.4)
Prior manic
episodes 1 to 10: 48.9% 11 to 20:
13.3% > 20: 36.6%
Prior
depressive episodes 0: 4.9% 1 to 10:
44.7% > 10:
48.8%
61% had at least one previous
hospitalization 18% hospitalized for the index episode | 4758/--/571/372 | 256 withdrew / Number lost to follow-up not
reported / 369 analyzed | Divalproex vs. Lithium vs.
Placebo
Time to 50% relapse of any mood episode
(95% CI), d: 275 (167 to not calculable
[NC]) vs. 189 (88 to NC) vs. 173 (101 to
NC)
Time to 25% relapse with mania
(95% CI), d: >365 (NC) vs. 293 (71 to NC) vs. 189 (84 to
NC)Time to 25% relapse with depression (95% CI), d:
126 (100 to 204) vs. 81 (33 to 234) vs. 101 (55 to 190) (p =
0.08 for divalproex vs. lithium) | Proportion of patients remaining in study
(estimated from Kaplan- Meier survival curve at 52 wk): 0.48 vs. 0.42 vs.
0.41 (p = 0.06)
Median time to 50%
survival without any mood episode based on 4-wk intervals, wk: 40 vs. 24 vs.
28 (no statistical analyses) | Mean changes from baseline in scores (Center
Effects model) MRS: 3.1 vs. 3.0 vs. 3.4 (p > 0.05 for all
analyses) DSS: 3.9 vs. 5.7 vs. 6.1 (p > 0.05 for all analyses) GAS:
−4.7 vs. −7.8 vs. −5.7 (p > 0.05
for all analyses)
Mean changes from baseline in scores (Mania
Subtype model) MRS: 1.7 vs. 2.6 vs. 2.7 (p > 0.05 for all
analyses) DSS: 3.6 vs. 7.0 vs. 4.4 (p < 0.001 Divalproex vs. Lithium;
p=0.02 Lithium vs. Placebo) GAS: −4.7 vs.
−10.8 vs. −6.2 (p=0.001 Divalproex vs.
Lithium; p=0.03 Lithium vs. Placebo) | | Not reported | Rate of AEs higher on... Divalproex than
Lithium: sedation, infection, tinnitus Lithium than Divalproex:
polyuria, thirst Divalproex than Placebo: tremor, weight
gain Lithium than Placebo: tremor
Divalproex vs.
Placebo Change in platelet count, 109/l: −53 vs. 3.4 (p
< 0.001) Change in white blood cell count, 109/l:
−1.1 vs. −0.3 (p <
0.009)
Change in hepatic enzymes: NSD | Open-label phase Total withdrawals: 199/571
(34.9%) Withdrawals due to adverse events: 10/199
(5.0%)
Divalproex vs. Lithium
vs. Placebo Double-blind phase Total withdrawals:
116/187 (62%) vs. 69/91 (76%) vs. 71/94
(75%) (p = 0.03 Divalproex < Lithium)
Withdrawals due to intolerance or noncompliance: 41/187 (22%),
32/91 (35%) vs. 11/94 (12%) (p=0.02
Divalproex < Lithium) | Fewer patients randomized to lithium than
divalproex. Failure to achieve remission within 3 months of manic episode
was a major reason for exclusion from randomization (28 (14.1%)
of 199 patients not randomized to maintenance phase). Study had inadequate
power to detect treatment differences in the primary outcome variable (i.e.,
0.3 instead of the planned power of > 0.8). High dropout rate may
have biased the results. Further data available in Commentary by
Baldessarini, 2000 and systematic review by Macritchie 2004. |
Gyulai, 2003 U.S. (Fair) | Same as Bowden, 2000; presents additional analyses
to Bowden, 2000 Outpatient setting implied | Same as Bowden, 2000 | Same as Bowden, 2000 | Same as Bowden, 2000 | Lorazepam, haloperidol, sertraline,
paroxetine | DSS and MRS for symptom severity (from SADS-C);
frequency unclear (weekly x 6 wk, biweekly till wk 12, then
monthly?).
Breakthrough depression was defined by either need
for antidepressant treatment, which should have been initiated if DSS
score> / = 25, or early discontinuation for depression,
including SADS-C suicide item score >/= 4, attempted
suicide, or hospitalization for depression. | Age, mean (SD), y: 39.2 (11.8) Male / Female: Data
not reported Ethnicity not reported | Same as Bowden, 2000 | 4758/-/571/372 (number screened from Baldessarini
2000) | 256/372 (68.8%) withdrew / Number lost
to follow-up not reported / 372 analyzed | Divalproex (N = 187) vs. Lithium(N
= 91) vs. Placebo (N = 94)
Early
Discontinuation for Breakthrough Depression: 12 (6%) vs. 9
(10%) vs. 15 (16%) (NSD for divalproex vs. lithium
and lithium vs. placebo; p = 0.017 for divalproex vs.
placebo) --Hospitalization for depression: 3 (1.6%) vs. 2
(2.2%) vs. 6 (6.4%) --Suicide attempt: 2 vs.
2 vs. 2
Early discontinuation for any reason: 116
(62%) vs. 69 (76%) vs. 71 (75%) (p
= 0.05) Among SSRI users: 23/41 (56%)
divalproex vs. 17/20 (85%) placebo (p = 0.043) | Predictors of Early Discontinuation for
Depression Negative Predictors: --Divalproex (OR
= 0.426 (0.182 to 0.997--interval not defined) vs. placebo; p
= 0.049)
Positive Predictors: --Higher
number of previous depressive episodes (OR = 1.30
[1.055 to 1.598] per category (p =
0.014) --Psychiatric hospitalizations (OR = 1.68
[1.100 to 2.577] per category (p =
0.017) | Time to Depressive Relapse: NSD (data not
reported) For the subset of open- label divalproex responders (n
= 142), time to depressive relapse was longer with divalproex (n
= 71) than lithium (n = 41) (p =
0.03). | Predictors of Depressive Relapse
Positive Predictors: --Higher lifetime number of manic and depressive
episodes (increase in OR = 1.12 [1.04 to
1.21] for every category increase; p =
0.002) --Female gender (OR = 1.98 [1.22 to
3.22]; p = 0.006 vs.
males)
Predictors of Worsening Depressive
Symptoms Positive Predictors: --Lifetime number of manic
episodes (p = 0.015) --Number of psychiatric
hospitalizations (p = 0.015) Negative
Predictors: --Baseline DSS score (p = 0.002) | Not reported (see Bowden, 2000) | Not reported (see Bowden, 2000) | Total withdrawals was reported as an efficacy
outcome measure (Early Discontinuation for Any Reason) Withdrawals
due to adverse events: Not reported (see Bowden, 2000) | Subgroup of SSRI-treated patients was analyzed
post hoc . This was the first study to suggest that the
life time number of manic episode is associated with continuing depressive
morbidity in bipolar disorder. Low placebo relapse rate reduced the
effect size, thereby decreasing the probability of detecting differences
between active treatment groups and the placebo group. |
Tohen, 2002 U.S. (Fair) | Multicenter Double-blind RCT (test of
noninferiority) Inpatient for at least one week then outpatient | Age 18 to 75 y; diagnosis of bipolar I disorder
(DSM-IV criteria), manic or mixed episode, with or without psychotic
features; Young Mania Rating Scale minimum total score of 20 | Olanzapine 5 to 20 mg/d vs. Divalproex 500 to 2500
mg/d for 3 wk | None | Lorazepam < 2 mg/d and not within 8 h of a
symptom rating scale; benztropine < 2 mg/d | Young Mania Rating Scale (YMRS, 11-item) and
Hamilton Depression Rating Scale (HDRS, 21-item) daily for one week then
weekly Response defined as >/= 50%
reduction in YMRS score Remission defined as end point YMRS
</= 12 | Olanzapine vs. Divalproex Mean (SD) age:
40.0 (12.1) vs. 41.1 (12.3) 42.6% male, 57.4% female
80.9% Caucasian | Nonpsychotic 54.6% Mixed
Episode 43.0% Manic Episode
57.0% Rapid Cycling 57.4% | 330/--/--/251 | 79/ Not reported /248 | Divalproex vs. Olanzapine Total YMRS score,
mean change from baseline (Primary Efficacy Variable): -10.4 vs.
-13.4 Lower limit of 95.76% one-tailed CI for assessment
of noninferiority: 0.96 (exceeds predefined −1.9 margin of
therapeutic equivalence) Difference in mean change in YMRS score: 3.0
(p < 0.03) | Responders: 42.3% vs.
54.4% (p = 0.058) Remission:
34.1% vs. 47.2% (p < 0.04) HDRS, mean
change from baseline: −3.46 vs. −4.92 (NSD) | Time to response: Faster on olanzepine (data not
reported) Time to remission, d (25th percentile): 6 vs. 3 Mean
change in YMRS score in subgroup... --without psychosis:
−8.7 vs. −14.1 (difference: 5.4; p <
0.001) --with psychosis: −12.8 vs. −12.6 (p
= 0.93) | | Monitored | Common ( > 10%)
treatment-emergent AEs: More common on olanzapine: Dry mouth,
increased appetite, somnolence More common on divalproex:
Nausea Greater weight gain on olanzapine (2.5 kg) vs. divalproex (0.9
kg) | Total withdrawals: 39/125 (31.2%) vs.
37/126 (35.7%) Withdrawals due to adverse events: 9
(7.1%) vs. 12 (9.6%); p = 0.50 | 3 Divalproex patients excluded from primary
efficacy analysis because of no postbaseline assessment. |
Tohen, 2003 U.S. (Fair) | Multicenter 47-wk double-blind
RCT Extension phase to study by Tohen, 2002 Tested for
noninferiority Inpatient for at least one week then outpatient | Same as Tohen, 2002 | Olanzapine 5 to 20 mg/d vs. Divalproex 500 to
2500mg/d for 47 wk | None | Same as Tohen, 2002 | Young Mania Rating Scale (YMRS, 11-item), Hamilton
Depression Rating Scale (HDRS, 21 item), Clinical global Impression scale
for bipolar disorder (CGI-BP) severity of illness rating, and Positive and
Negative Syndrome Scale (PNSS) daily for one week then weekly from weeks 1
to 5, biweekly from weeks 5 to 11, monthly from weeks 11 to 23, and
bimonthly from weeks 23 to 47
Definitions Symptomatic
remission of mania: YMRS </= 12. Symptomatic
remission of mania and depression: endpoint total YMRS
</= 12 and HDRS </=
8. Syndromal remission of mania: no "A" criterion worse than mild in
severity and no more than two "B" criteria rated as mild in severity using
DSM-IV criteria Syndromal remission of mania and depression was
defined as the preceding mania criteria plus the following depression
criteria: no DSM-IV A criteria for a major depressive episode that were
worse than mild in severity and the presence of no more than three A
criteria rated as mild Symptomatic relapse into an affective episode
(depression mania or mixed): | Olanzapine vs. Divalproex Mean (SD) age:
40.0 (12.1) vs. 41.1 (12.3) 42.6% male, 57.4%
female 80.9% Caucasian | Mean (SD) YMRS total score: 27.7 (5.9;
severe) Mixed bipolar 43.0% Rapid cycling
57.4% Psychotic 45.4% Treatment
resistant (did not respond to previous adequate treatment for acute mania
with lithium, valproate, or carbamazepine) 21.1% | --/--/251/251 | 187 / 25 / 248 | Divalproex vs. Olanzapine YMRS total score,
mean difference: 2.4 (p = 0.002) Mean change in YMRS
total score(baseline to wk 47): −12.5 vs. −15.4 (p
= 0.03) Improvement in YMRS was significantly superior
from wk 2 to 15and wk 23; NSD from wk 30 to 47. NSD in HDRS, PNSS,
and CGI-BP severity of illness | Median time to symptomatic / syndromal remission
of mania,d: 62 / 109 vs. 14 / 28 (p = 0.05 / p =
0.01) Symptomatic mania remission rates: 45.5% vs.
56.8% (p=0.10) Syndromal mania remission
rates: 38.2% vs. 50.8%
(p=0.06) Time to symptomatic / syndromal remission of
both mania and depression (25th percentile),d: 13 / 34 vs. 14 / 7
[sic ] (p = 0.62 /
p = 0.86) p = 0.86 / p =
0.62
Symptomatic remission of both mania and depression:
30.9% vs. 30.9% (p =
1.00) Syndromal remission of both mania and depression:
27.6% vs. 29.8% (p=0.78) | Time to symptomatic recurrence of any affective
episode (25th percentile),d: 27 vs. 27 Symptomatic recurrence of any
affective episode:13/23 (56.5%) vs. 14/33 (42.4%) (p
= 0.42) Time to syndromal recurrence of any affective
episode (median),d: 42 vs.14 Syndromal recurrence of any affective
episode: 13/20 (65.0%) vs. 20/31 (64.5%) (p
= 1.00) | Relation of valproate serum concentration to
outcome (data not shown here): NSD for any analyses | Monitored | Treatment-emergent
AEs
Significantly more common on olanzapine:
somnolence, dry mouth , increased appetite, weight gain, akathisia,
increased alanine aminotransferase
Significantly more common
on divalproex: nausea, nervousness, rectal disorder, low albumin, low
platelets
Olanzapine vs. divalproex Mean weight gain:
2.79 vs. 1.22 kg (p = 0.001) Mean change in cholesterol:
9.7 vs. −2.33 mg/dl (p = 0.007) Mean change
in Fridericia-corrected QT interval: 7.97 msec vs. −3.06 (p
= 0.002) Potentially clinically significant change in QTc
interval (> 430 in men, >450 in women): 2/102
(2.0%) vs. 2/96 (2.1%) (p = 1.00) | Olanzapine vs. Divalproex
Total
withdrawals: 106/125(84.8%) vs. 106/126 (84.1%) (p
= 1.00)
Withdrawals due to adverse events: 31/125
(24.8%) vs. 25/126 (19.8%) (p =
0.37)
Withdrawals due to weight gain: 4/125
(3.2%) vs. 0/126 (0.0%) | High dropout rate limits the power to detect
differences in relapse. For most patients, initial olanzapine doses
(15 mg/d) may be therapeutic while initial divalproex doses (750 mg/d)may be
subtherapeutic. This difference may have favored an earlier response with
olanzapine. |
Zajecka, 2002 U.S. (Fair) | Multicenter, double-blind, double-dummy,
parallel-group RCT Inpatient (< 3 wk) then outpatient (9 wk)
setting | Randomization criteria: Age 18 to 65 y; bipolar
disorder type I (DSM-IV); hospitalized for an acute manic episode (defined
as a score of >/= 25 on the Schedule for Affective
Disorders and Schizophrenia-Change Version (SADS-C) Mania Rating Scale
(MRS), with at least 4scale items rated >/=
3).
Improvement criteria (on or before day 21, for discharge
from hospital and follow-up as outpatients for remainder of study): SADS-C
MRS score reduced >/= 30% from the last day
of screening, with no SADS-C item score > 3, and discharge
recommended by the investigator. | Divalproex Delayed-release starting at 20mg/kg/d
and titrated to a maximum of 20 mg/kg/d + 1000 mg (range, 750 to
3250 mg) vs. Olanzapine 5to 25 mg/d for 12 wk | 1- to 3-day non-drug run-in 1- to 3-day
washout of previous psychoactive medications | Lorazepam, benztropine, chloral hydrate, zolpidem
(but not within 8 h prior to efficacy ratings) | MRS at baseline, and days 3, 5, 7, 10, 14, 21, 28,
42, 56, 70, and 84; Brief Psychiatric Rating Scale (BPRS) at baseline and
days 3, 5, 7, 14, 21, 28, 42, 56, 70, and 84; Hamilton Rating Scale for
Depression (HAM-D) at baseline and days 7, 14, 21, 28, 42, 56, 70, and 84;
Clinical Global Impressions-Part I, severity of illness scale (CGI-S) at
baseline, and days 3, 7, 14, 21, 28, 42, 56, 70, and 84 | Divalproex (N = 63) vs. Olanzapine (N
= 57) Age, mean (SD), y: 38.9 (12.1) vs.38.1
(12.2) Male / Female: 56% / 44% vs.
53% / 47% Ethnicity, n
(%) --Asian/Pacific Islander: 2 (3) vs. 1
(2) --White: 50 (79) vs. 40 (70) --Black: 8 (13) vs. 14
(25) --Other: 3 (5) vs. 2 (4) | DSM-IV diagnosis Mixed mania: 31
(49%) vs. 26 (46%) Rapid cycling: 19
(30%) vs. 16 (28%) | Numbers screened, eligible, enrolled not reported
/ 120randomized | 67 / 16 / 115 | Divalproex vs. Olanzapine
Change
from baseline to day 21 (last observation carried forward), mean MRS
(with baseline as covariate, Primary Efficacy Variable): −14.9
vs. −16.6 (NSD) BPRS: −8.1 vs.
−10.2 (NSD) HAM-D: −6.7 vs. −8.1
(NSD) CGI-S: −0.8 vs. −1.0 (NSD) | NSD in antipsychotic effect (although numbers
small and variability of change in BPRS scores was
high).
Data for 12-wk tx were not reported. | | | Monitored | Divalproex (N = 61) vs. Olanzapine (N
= 57) Increase in weight (baseline to final evaluation),
mean, kg: 2.5 vs. 4.0 (p = 0.049)
Divalproex (N
= 63) vs. Olanzapine (N = 57) Adverse
Events Significantly more frequent on olanzapine than divalproex:
somnolence (29% vs. 47%), weight gain, rhinitis,
edema, speech disorder (slurred speech) Significantly more frequent
on divalproex: None
Deaths and Serious Adverse
Events 1 Death on olanzapine attributed to diabetic ketoacidosis that
was considered to be possibly/probably related to study drug 5
Divalproex patients: abnormal electrocardiogram results; anticholinergic
syndrome; catatonic reaction; psychotic depression; somnolence
(possibly/probably related to study drug) 2 Olanzapine patients:
depression, diabetic ketoacidosis (possibly/probably related to study
drug)
Change from baseline to final values, mean Total
cholesterol, mg/dl: −1.69 vs. 13.29 (p =
0.019) LDL, mg/dl: −4.43 vs. 8.78 (p =
0.022) Platelet count (x 109/l): 52 19 vs 0 78 (p < 0
001) | Divalproex vs. Olanzapine Total
withdrawals: 45/63 (71%) vs. 38/57
(67%) Withdrawals due to adverse events: 7/63
(11%) vs. 5/57 (9%) p = 0.766 | Washout period of 1 to 3 days may be inadequate.
Baseline MRS scores were significantly different; effect on results was not
explained. This trial used higher doses of divalproex and serum
concentrations were also higher than those in the trial by Tohen. The higher
doses would not intuitively explain the difference in results between
Tohen's positive study and this negative study. Limited by
selection bias, as previous study drug failures were excluded. |
Bowden, 2003 Australia, Canada, Greece, New
Zealand, U.K., U.S., Yugoslavia Lamictal 606 Study (Fair) | Multicenter double-blind, parallel-group, placebo-
controlled RCT with 2-wk screening phase, 8- to 16-wk open-label phase on
lamotrigine treatment, and a 76-wk double-blind phase Clinic setting | 18 yr or older; bipolar I disorder; manic or
hypomanic (DSM-IV) currently or within 60 d; manic or hypomanic symptoms at
enrollment; at least 1 additional manic or hypomanic episode and 1 depressed
episode within 3 yr of enrollment; Clinical Global Impression-Severity
(CGI-S) score of 3 or less for at least 4 continuous wk during open-label
phase | Open-label: Lamotrigine 100 to 200 mg/d for 8 to
16 wk Double-blind: Lamotrigine 100 to 400 mg/d vs. Lithium
titrated to serum concentrations 0.8 to 1.1 mEq/l vs. Placebo for up to 76
wk | Run-in: beginning at wk 8 of open-label
lamotrigine, patients who had reached a stable dose of lamotrigine and met
criterion for response (CGI- S scale score of 3 or less for at least 4
continuous wk) were eligible for double-blind phase. Patients who
developed adverse events were not randomized. Patients who did not meet
response criteria by wk 16 were discontinued from study. | Open-label phase: AEDs, psychotropic medications
up to 1 to 2 wk before entry into double-blind
phase.
Double-blind phase: No psychotropics except
short-term, intermittent use of chloral hydrate, lorazepam, temazepam, or
oxazepam at low doses. Institution of antidepressant, antipsychotic,
benzodiazepine, AED, mood stabilizer, and electroconvulsive therapy for a
mood episode constituted the primary study end point. | Time to intervention (addition of pharmacotherapy
or electroconvulsive therapy) for any mood episode (primary efficacy end
point); time to early discontinuation for any reason; time to intervention
for manic, hypomanic, or mixed episode; time to intervention for depressive
episode; scores on Mania Rating Scale (MRS), Hamilton Rating Scale for
Depression (HAM-D, 17-item), Clinical Global Impression- Severity (CGI-S)
and -Improvement (CGI-I), and Global Assessment Scale (GAS) weekly for 4 wk,
biweekly through wk 8, then every 4 wk through wk 76. | Open-label Lamotrigine; Double-blind Lamotrigine,
Lithium, and Placebo Mean (SD) age: 40.7 (11.8); 40.6 (12.6), 41.9
(11.3) vs. 40.9 (11.0) Male: 50%; 45%,
48% vs. 49% Ethnicity not reported | Open-label Lamotrigine; Double-blind Lamotrigine,
Lithium, and Placebo Mean (SD) MRS: 22.9 (6.7); 22.3 (6.8), 22.3
(5.6) vs. 22.4 (7.8) History of psychotic episodes: 46%;
38%, 46% vs. 41% Ever
hospitalized for mood-related disturbance: 66%; 60%,
67% vs. 61% Ever attempted suicide:
29%; 28%, 41%, 19% (Lithium
vs. Placebo, p=0.01) | --/--/349/175 | Open-label phase (N=349): 135/30/184
(completed)
Double-blind phase: 41/5/171 | Lamotrigine vs. Lithium vs. Placebo (p-values
shown for lamotrigine vs. lithium, lamotrigine vs. placebo, and lithium vs.
placebo, respectively)
Median time to any mood episode
(95% CI), d: 141 (71 to > 547) vs. 292 (123 to >
547) vs. 85 (37 to 121) (p = 0.46, 0.02, and
0.003)
Median survival in study (95% CI), d: 85
(44 to 142) vs. 101 (59 to 202) vs. 58 (34 to 108) (p = 0.72,
0.03, and 0.07)
Proportion of patients remaining in study
(estimated from Kaplan-Meier survival curve at 76 wk, Figure 1 of article): 0.43 vs. 0.47 vs. 0.15 (p
= 0.46, 0.02, and 0.003) | Time to mania and depression episodes: Not
evaluable for lamotrigine and lithium; 269 (95% CI: 183 to
> 547) for placebo
Kaplan-Meier survival estimates to
manic episode (from Fig. 2 of
article): 0.65 vs. 0.55 vs. 0.40 (p = 0.09, 0.28,
0.006)
Kaplan-Meier survival estimates to depressive episode
(from Fig. 2 of article): 0.80
vs. 0.70 vs. 0.40 (p=0.36, 0.02, 0.17) | Mean change from baseline scores; calculated
differences and p-values shown for lamotrigine vs. lithium, lamotrigine vs.
placebo, and lithium vs. placebo
MRS: 1.79 vs.
−0.04 vs. 2.3; calculated differences: 1.83, −0.51,
and −2.34 (p = 0.03, p > 0.05, and p
= 0.001)
HAM-D: 2.05 vs. 2.68 vs. 3.92;
calculated differences: −0.63, −1.87, and
−1.24 (p > 0.05, p = 0.03, and p >
0.05)
GAS: −3.19 vs. −3.85 vs.
− 5.63; calculated differences: 0.66, 2.44, and 1.78 (p >
0.05 for all comparisons)
CGI-S: 0.37 vs. 0.44 vs. 0.56;
calculated differences: −0.07, −0.19, and
−0.12 (p > 0.05 for all comparisons) | | Monitored | Lamotrigine vs. Lithium vs. Placebo Adverse
events occurring in at least 10% of patients and at rates
showing treatment differences
--Headache: 12/59
(20%) vs. 2/46 (4%) vs. 11/69 (6%) (p
= 0.02, lamotrigine vs. lithium)
--Diarrhea: 3/59
(5%) vs. 13/46 (28%) vs. 6/69 (9%) (p
= 0.002, lamotrigine vs. lithium; p = 0.009, lithium
vs. placebo
Other common AEs (no treatment differences): Any
rash, infection, somnolence, nausea, insomnia, influenza | Lamotrigine vs. Lithium vs.
Placebo
Total withdrawals: 13 (22.0%) vs. 18
(39.1%) vs. 10 (14.3%)
Withdrawals
due to adverse events: 3 (5%) vs. 11 (24%) vs. 3
(4%) (p = 0.01 for both lithium vs. lamotrigine and
lithium vs. placebo) | Slow rate of recruitment led to closure of lithium
arm about midway through study and termination of study before full planned
enrollment (100 per group). Possible implications of baseline differences in
suicide rates on study results were not reported. Higher enrollment of
patients with more severe depression (higher rate of past suicide attempts)
in the lithium group may have influenced treatment results for depressive
episodes. Double-blind results are confounded by discontinuation of patients
who experienced AEs or lack of efficacy to lamotrigine in open- label phase.
Survival in study, in which all dropouts were included as events, was used
to confirm the primary efficacy analysis, which excluded dropouts other than
those due to defined events. |
Calabrese, 2003 U.S., Canada, Denmark,
Finland, U.K. Lamictal 605 Study (Fair) | Multicenter, double-blind, double-dummy, placebo-
controlled, parallel-group RCT with open-label run-in
phase Outpatient clinic setting | Age at least 18 y; bipolar I disorder; currently
experiencing a major depressive episode (DSM-IV) or residual depressive
symptoms present from a major depressive episode within 60 d of screening;
at least 1 manic or hypomanic episode within 3 y of enrollment; at least 1
additional depressed episode (including a mixed episode) within 3 y of
enrollment. | Open-label phase: Lamotrigine titrated to 100 to
200 mg/d as adjunctive or monotherapy for 8 to 16 wk (target dose halved
when used adjunctively with valproate)
Double-blind phase:
Lamotrigine 50 mg/d vs. Lamotrigine 200 mg/d vs. Lamotrigine 400 mg/d vs.
Lithium titrated to serum concentrations of 0.8 to 1.1 mEq/l vs. Placebo for
76 wk | 8- to 16-wk open-label run- in phase on
lamotrigine monotherapy or adjunctive therapy (target dose, 100 to 200
mg/d); beginning at wk 8 of the open-label phase, patients who had Clinical
Global Impression- Severity of Illness (CGI-S) scores of 3 (mildly ill) or
lower maintained for at least 4 continuous wk were randomized. 1- to
2-wk washout of previous psychotropic medications including AEDs; 4-wk
washout for fluoxetine | Chloral hydrate, lorazepam, temazepam, oxazepam,
midazolam | Time to intervention (addition of pharmacotherapy
or electroconvulsive therapy) for any mood episode (primary efficacy end
point); time to intervention for a manic or hypomanic episode; time to
intervention for a depressive episode; HAM- D, MRS, CGI-S, and Global
Assessment Scale (GAS), at baseline (day 1 of double- blind phase) and
during double-blind phase (intervals not reported). | Open-label lamotrigine (N = 958),
Placebo (N = 121), Lithium (N = 120)
vs. Lamotrigine (N = 169) Age, mean (SD), y: 42.2
(12.2) vs. 42.1 (13.0) vs. 43.6 (12.3) vs. 44.1
(11.7) Men: 39% vs. 50%
vs. 40% vs. 41% Ethnicity not
reported | History of psychotic episodes: 31% vs.
30% vs. 29% vs. 29% Ever
hospitalized for mood-related distrubances: 66% vs.
64% vs. 63% vs. 57% Ever
attempted suicide: 37% vs. 36% vs. 35%
vs. 25% Age at first depression, mean (SD), y: 22.7
(11.6) vs. 22.4 (11.9) vs. 23.1 (12.1) vs. 23.5 (11.8) Age at first
mania/mixed episode, mean (SD), y: 26.7 (12.5) vs. 25.7 (12.8) vs. 28.4
(14.6) vs. 27.7 (12.2) 4 to 6 mood episodes in past year:
28% vs. 34% vs. 32% vs.
25% | Number screened not reported / 966 eligible for
open- label phase, 480 eligible for double- blind phase / Number enrolled
not reported / 463 randomized | Open-label phase: 486/966 (50.0%)
withdrew; 60/966 (6%) were lost to follow-up from the open-label
phase Double-blind phase: 156/463 (33.7%) withdrew / 25/463
(5.4%) lost to follow-up / 457 analyzed | Lamotrigine 200/400 (N = 165) vs.
Lithium (N = 120) vs. Placebo (N = 119); p-values
shown for lamotrigine vs. lithium, lamotrigine vs. placebo, and lithium vs.
placebo
Time to any mood episode (primary efficacy measure),
median (95% CI), d: 200 (146 to 399) vs. 170 (105 to not
evaluable) vs. 93 (58 to 180); p = 0.915, p = 0.029,
and p = 0.029
Overall survival in study, median
(95% CI), d: 92 (59 to 144) vs. 86 (63 to 111) vs. 46 (30 to
73); p = 0.516, p = 0.003, and p =
0.022
Proportion of patients remaining in study for time to
intervention for any mood episode at 76 wk (estimated from Kaplan-Meier
survival curve, Fig. 2A): 0.36
vs. 0.40 vs. 0.25; p = 0.915, 0.029, and 0.029 | Calculated differences and p- values shown for
lamotrigine vs. lithium, lamotrigine vs. placebo, and lithium vs.
placebo
Intervention-free for depression at 1 y:
57% vs. 46% vs. 45%; calculated
differences: 11%, 12%, and 1% (p
= 0.434, p = 0.047, and p =
0.209)
Intervention-free for mania at 1 y: 77%
vs. 86% vs. 72%; calculated differences:
-9%, 5%, and 14% (p = 0.125,
p = 0.339, and p = 0.026) | Change from baseline, mean; calculated differences
and p-values shown for lamotrigine vs. lithium, lamotrigine vs. placebo, and
lithium vs. placebo
HAM-D (17-item): 2.5 vs. 2.9 vs. 4.9 (p
> 0.05, p < 0.05, p < 0.05)
MRS: 0.7 vs.
0.7 vs. 1.1 (p > 0.05 for all comparisons)
GAS:
−2.8 vs. −4.1 vs. −6.9 (p > 0.05,
p < 0.05, p < 0.05) | Change from baseline, mean CGI-Severity of
Illness: 0.7 vs. 0.4 vs. 0.3; p < 0.05 lithium or lamotrigine vs.
placebo CGI-Improvement: 2.6 vs. 2.5 vs. 2.5 (NSD) | Not reported | Open-label phase (N = 958), Placebo (N
= 121), Lithium (N = 120), vs. Lamotrigine (N
= 169)
Most common treatment-emergent adverse
events showing treatment differences, n (%) Any rash: 104
(11) vs. 3 (2) vs. 5 (4) vs. 12 (7); p < 0.05 lamotrigine
vs. placebo Somnolence: 83 (9) vs. 7 (6) vs. 16 (13) vs. 16
(9); p < 0.05 lithium vs. placebo Diarrhea: 81 (8) vs. 10 (8)
vs. 19 (16) vs. 12 (7); p < 0.05 lamotrigine vs.
lithium Tremor: 46 (5) vs. 6 (5) vs. 20 (17) vs. 9 (5); p <
0.05 lithium vs. placebo and lamotrigine vs. lithium | Double-blind phase Placebo (N =
121) vs. Lithium (N = 121) vs. Lamotrigine (N
= 221) Total withdrawals: 43 (36%) vs. 45
(37%) vs. 68 (31%) Withdrawals due to adverse
events: 15/169 (9% ) vs. 19/120 (16% ) vs. 12/121
(10%) (NSD) | An a priori decision was made to combine the
existing 200- and 400-mg/d lamotrigine groups for the primary analysis of
efficacy. Survival in study, in which all dropouts were included as events,
was used to confirm the primary efficacy analysis, which excluded dropouts
other than those due to defined events.
Efficacy and safety
comparisons between lamotrigine and lithium are limited because patients
with intolerance or lack of efficacy to open-label lamotrigine were excluded
from the maintenance phase. Even with the enriched enrollment of lamotrigine
responders, there was no significant difference between lamotrigine and
lithium for the primary efficacy measure (time to any mood episode). |
McIntyre, 2002 Canada (Poor) | Single-blind,
parallel-group RCT Bipolar Clinic setting | Bipolar I/II disorder (DSM-IV) with most recent
episode depression. Patients receiving divalproex or lithium must
have received the medication for at least 2 wk. | Topiramate 50 to 300 mg/d (mean dose: 176 mg/d)
vs. Bupropion sustained release (SR) 100 to 400 mg/d (mean dose: 250 mg/d)
(added on to mood stabilizer) for 8 wk | None | Atypical antipsychotics, lithium (mean
+/− SD dose: 980 +/− 388.3
mg/d; mean plasma concentration: 1.16 mEq/l; mean duration: 4.4 y),
divalproex (1106 +/− 400.36 mg/d; 498.4 mol/l; 6.2
y) | Hamilton Depression Rating Scale (HDRS-17 item);
Young Mania Rating Scale (YMRS); Clinical Global Impression for Severity
(CGI- S) and Improvement (CGI-I); and AMDP [not
defined] side effects rating scale, at baseline and
weekly. Montgomery Asberg Depression Rating Scale (MADRS) at baseline
and end point. Primary efficacy measure was percentage of patients
responding. Response was defined a priori as >/=
50% decrease from baseline in the mean total HDRS-17
score. Remission was defined as an end point HDRS 17 score
</= 7. | Topiramate (N = 18) vs. Bupropion SR
(N = 18) Age, mean, y: 39 vs. 43 Male/Female: 11/7
vs. 10/8 | Age of onset of illness, mean, y: 24 vs.
22 Rapid cyclers: 8 (44%) vs. 7
(39%) Number of lifetime episodes, mean --Manic:
4.3 vs. 3.0 --Hypomanic: 1.8 vs. 2.4 --Depressive: 4.0 vs.
3.0 Duration of current episode, mean, mo: 6.5 vs.
7.5 Concomitant psychiatric medication, n --Atypical
antipsychotics: 3 vs. 3 --Lithium: 5 vs. 8 --Divalproex: 13
vs. 10 Previously treated with benzodiazepines: 29% vs.
35% Previously treated with antidepressants:
40% vs 45% | Numbers screened and eligible not reported/36
enrolled/36 randomized | 13/36 (36.1%) withdrew/None lost to
follow-up/36 analyzed | Responder rate: 56.2% vs.
58.7% (p-value not reported) Calculated difference in
responder rate: −2.5%
Remission rate:
24.8% vs. 27.5% Calculated difference in
remission rate: −2.7%
Time to
response: 2 to 4 wk for both treatment groups | Mean HDRS-17 scores, calculated change from
baseline to 8 wk : 10.5 vs. 10.5 (NSD) | CGI-I scores: NSD (data not reported) CGI-S
scores: Not reported Mean YMRS scores, calculated change from
baseline to end point: −5 vs. −6 (NSD) | | Monitored | Topiramate vs. Bupropion SR Adverse event
rate: 11/18 (61.1%) vs. 9/18
(50.0%)
Topiramate (n = 14) vs.
Bupropion SR (n = 13)
Most common adverse events
reported more frequently on Bupropion Difficulty sleeping:
16.0% vs. 27.8% (p =
0.03) Paresthesias: 17.4% vs. 27.6%
(NSD) Tremors: 18.1% vs. 25.1%
(NSD)
Mean weight loss, kg: 5.8 vs. 1.2 (p =
0.04)
No patient exhibited a manic switch | Topiramate vs. Bupropion Total withdrawals:
8/18 (44.4%) vs. 5/18 (27.8%) Withdrawals due
to adverse events: 6/18 (33.3%) vs. 4/18
(22.2%) | Lacked placebo arm. Small sample size; lacked
sufficient power to detect a treatment difference. Concomitant medications
confound results. Results should be considered preliminary. |
Okuma, 1990 Japan (Poor) | Multicenter, double-blind, double-dummy
RCT Outpatient and inpatient psychiatric university clinics and
hospitals | Endogenous manics (ICD-9); also met criteria for
bipolar disorders in the affective disorders of DSM-III; psychopharmacologic
treatment- naïve or experienced; age 13 to 65 y | Carbamazepine starting at 400 mg/d and titrated to
symptoms and adverse effects Lithium starting at 400 mg/d and
titrated to symptoms and adverse effects for 4 wk | None | Antipsychotics without sufficient antimanic effect
prior to study could be continued at stable doses | 5-point severity of illness scale (ranging from
Normal to Extremely Severe) at baseline and weekly; 6-point scale for global
improvement rate relative to first day of treatment (ranging from Markedly
Improved to Alteration to Depressive or Mixed State), recorded weekly;
6-point scale for Final Global Improvement Rate (FGIR) on last day of
treatment; 14-item Clinical Psychopharmacology Research Group (CPRG) Rating
Scale for Mania, Doctor's Use, before and weekly | Carbamazepine (N = 50) vs. Lithium (N
= 51) Age, mode, y: 20 to 29 y (range, less than 19 to
over 70 y; note: this exceeds eligible age limit) Male/Female: 26/24
vs. 22/29 Ethnicity: not reported | Bipolar, Manic: 49 vs. 48 Bipolar, Mixed: 1
vs. 3
At least moderate severity: 43 (86.0%) vs.
44 (86.3%)
Inpatient: 47 (94.0%) vs.
40 (78.4%) Outpatient: 3 (6.0%) vs. 11
(21.6%) | Numbers screened and eligible not reported/105
enrolled/105 randomized | 24 withdrawn/3 lost to follow-up/101 analyzed | Carbamazepine vs. Lithium
Marked or
Moderate Global Improvement, final assessment: 62% vs.
59% (NSD) Marked or Moderate Global Improvement,
wk 1: 11/50 (22.0%) vs. 5/51 (9.8%) | Total CPRG scores for mania, wk 4: 35.3 vs. 39.2
(NSD)
Serum carbamazepine concentration in good (N
= 20) vs. poor (N = 13) responders, wk 4: 8.0 vs.
6.3 mcg/ml (p < 0.05); NSD in daily doses
Serum
lithium concentration in good (N = 19) vs. poor (N =
9) responders: 0.41 vs. 0.56 mEq/l (p < 0.10); NSD in daily
doses | | | Monitored | Carbamazepine vs. Lithium
Frequency
of adverse events: 60% vs. 43%
(NSD)
Cutaneous symptoms (exanthema): 12% vs.
0% (p < 0.05) | Carbamazepine vs. Lithium
Total
withdrawals: 9/51 (17.6%) vs. 15/54
(27.8%)
Withdrawals due to adverse events: 5/51
(9.8%) vs. 0/54 (0.0%) (p < 0.05) | Quality of trial conduct is questionable; 2
lithium patients were given only placebo tablets of carbamazepine by mistake
and an erroneous report of blood concentration of lithium led to unblinding
of treatment in one case. Concomitant antipsychotics "without sufficient
antimanic effects" is unclear. Their use may have confounded the
results. |
Brown, 2006 USA
bipolar I
depression | RCT, multicenter, inpatient and outpatient | Inclusion- Total score 20 or greater on MADRS and
4 on CGI-S; at least 1 episode of mania or mixed to require treatment w/mood
stabilizer or antipsychotic Exclusion- Serious suicide risk;
substance abuse in last 30 days; YMRS 14 or more; currently taking or
previously failed on olanzapine or lamotrigine | olanzapine/fluoxetine combination (OFC) 6/25,
6/50, 12/25, or 12/50 mg/day vs. lamotrigine 200 mg/day 7 weeks | 2 day screening and all patients were tapered off
all meds24 hours prior to randomization/ | Anticholinergic meds allowed for EPS and
benzodiaepines or other hypnotics | Overall bipolar status as measured by CGI-S,
MADRS, YMRS, BPRS, CGI-I, PGI, GAF (Global Assessment of Functioning Scale),
MOS, BSI (Brief Symptom Inventory) Patients evaluated at
randomization, 3 days, 4 days, then weekly` | 37.0 years 60.0%
female 81.7 white | Age of onset 19.0 years Outpatients
99% | NR/NR/NR/410 | 139/47/410 | Change from baseline at 7 weeks- mixed effects
model repeated measures (SE) OFC vs. lamotrigine CGI-S
−1.43( 0.06) vs. −1.18(0.06) P =
0.002 MADRS −14.91(0.49) vs.
−12.92(0.50) P = 0.002 YMRS
−1.68(0.18) vs. −0.94(0.18) P=
0.001 GAF 11.00((0.52) vs. 9.22(0.52) P =
0.010 BSI −0.8 (0.05) vs.
−0.67(0.05) P= 0.028 CGI−I
2.41(0.06) vs. 2.63(0.06) P = 0.003 PGI 2.59(0.06)
vs. 2.84(0.06) P = 0.002 BPRS
−11.62(0.55) vs. −10.80(0.57) P =
0.253 | OFC vs. lamotrigine Responders (> 50
reduction in MADRS) 68.8% vs. 59.7% P =
0.073 Time to response median (95% CI) 17(14 to 22) vs.
23 (21 to 34) P = 0.010 | | | Evaluation of TEAEs, discontinuation due to Aes,
vital sign measurements and lab tests, all via MedDRA | Treatment-emergent AEs OFC vs. lamotrigine
% Smnolence 18.5 vs. 8.3 P =
0.003 Inceased appetite 17.6 vs. 8.3 P = 0.008 Dry
mouth 15.6 vs. 5.9 P = 0.002 Increased weight 14.1 vs.
2.0 P < 0.001 Dizziness 13.7 vs. 7.8 P =
0.078 Sedation 13.7 vs. 2.5 P < 0.001 Headache 11.7 vs.
9.3 P = 0.52 Tremor 10.7 vs. 1.5 P <
0.001 Fatigue 8.3 vs. 5.4 P = 0.328 Nausea 7.8 vs.
7.8 P = 0.99 Insomnia 4.4 vs. 8.8 P 0.076 Rash 2.9
vs. 6.9 P = 0.071 Suicidal or self-injurious
behavior 0.5 vs. 3.4 P = 0.037 | 139 withdrawals 32 due to AEs | |
Nierenberg,
2006 USA STEP-BD Bi-polar depression | RCT open-label | Inclusion - 1) were at least 18 years old, 2) met
criteria for bipolar disorder type I or II with a current DSM-IV major
depressive episode of at least 8 weeks before pathway entry, and 3) had not
responded to treatment in first 12 weeks of standard or randomized care
pathways for bipolar depression, or had a well-documented failure (e.g., a
medical chart was available) to respond to at least two trials of
antidepressants or an antidepressant and mood stabilizer regimen. Patients
were required to be taking a mood stabilizer or agree to begin treatment
with a mood stabilizer. Only patients who refused ECT at this stage were
eligible for randomization to the open-label treatment conditions Exclusion-
history of nonresponse to, intolerance of, or any medical contraindications
to at least two of the study medications; if they met criteria for mixed
episode or hypomania or if they met criteria for current substance abuse or
dependence diagnosis | Lamotrigine doses started at 50 mg/day for 2
weeks, followed by 50 mg b.i.d. for 2 weeks, then increases in daily dose
every week until the target dose of between 150 and 250 mg/day was
reached.
Inositol doses started at 2.5 to 5 g with a target
dose of between 10 and 25 g.
Risperidone doses started at
between 0.5 and 1.0 mg with titration up to 6 mg as
tolerated.
up to 16 weeks | None at this point | Yes- Subjects were managed with an optimized mood
stabilizer regimen (lithium, valproate, combined lithium and valproate, or
carbamazepine) plus either one or two antidepressants. | recovery rate within equipoise randomization
strata. Recovery was defined as 1) no more than two symptoms meeting DSM-IV
threshold criteria for a major depressive, manic, or hypomanic episode, as
determined with the clinician-administered Clinical Monitoring Form, and 2)
no significant symptoms present for 8 weeks, consistent with the DSM-IV
definition of full remission. Secondary outcome measures included
Clinical Global Impression (CGI) severity ratings, Clinical Monitoring Form
SUM-D and SUM-M scores, and Global Assessment of Functioning scores | 33.0 – 53.8 years median age
range 55% female 86% white | | 67/NR/NR/66 | NR/NR/69 (Three subjects were willing to accept
random assignment to any of the three treatments and therefore are included
in two strata and are counted twice in the pairwise comparisons) | Recovery rate lamotrigine 23.8% vs.
inositol 17.4% vs. risperidone 4.6%, SUM-D score
base/end lamotrigine 7.0/3.9 vs. inositol 7.7/6.6 vs. risperidone 6.3/7.6 a
significant difference in score between those assigned to lamotrigine and
those assigned to risperidone (normal approximation z=2.85,
p=0.004) or inositol (normal approximation
z=−2.14, p=0.03).
CGI
rating base/end lamotrigine 4.6/2.9 vs. inositol 4.2/3.9 vs. risperidone
4.4/4.1 a significant difference in score between those assigned to
lamotrigine and those assigned to risperidone (normal approximation
z=2.85, p=0.004) or inositol (normal approximation
z=−2.29, p=0.02). | Duration of treatment (weeks) lamotrigine 12.2 vs.
inositol 8.6% vs. risperidone 5.8 | | | NR | NR | NR NR | |
Nolen, 2007 USA and
Holland Refractory Bi Polar depression | RCT open-label | Inclusion - at least 18 years; met criteria for
DSM-IV bipolar I or II disorder, current major depressive episode Patients
had to be currently on treatment with a mood stabilizer at adequate plasma
level, all patients had previously been treated with a conventional
antidepressant (SSRI, TCA, venlafaxine or bupropion) in an adequate dose and
during at least 6 weeks without sufficient response or tolerance (including
switch into mania or hypomania) in the current or in a prior
episode. Exclusion- current alcohol or substance abuse or dependence;
a severe neurological or other somatic illness; (risk of) pregnancy; current
or recent (1 week) use of another antidepressant, an antipsychotic, or a
benzodiazepine other than lorazepam >4 mg/day. | tranylcypromine 100 mg/day vs. lamotrigine 400
mg/day
10 weeks | NR | currently on treatment with a mood stabilizer at
adequate plasma level | Clinical Global Impressions Scale for Bipolar
Illness (CGI-BP) the IDS-C and the Young Mania Rating Scale
(YMRS)
Assessments at baseline, week 1, week 2 and from then
biweekly until week 10 | 46.2 years 47.4% female
Ethnicity NR | Duration of illness 23.6 years | 1242/NR/NR/20/20 8/0/19 | response of depression tranylcypromine (5/8,
62.5%) lamotrigine (4/11, 36.4%) (P =
NS). | | Tranylcypromine vs. Lamotrigine, mean (SD) change
from baseline CGI-BP-Severity of depression 1.0 (2.8) vs. 0.6
(1.8) YMRS scores +1.25 (3.3) vs. +2.2
(7.9) CGI-BP-Severity of mania 0.0 (0.0) vs. 0.2
(1.0) Switch into mania (n, %) 0 (0) vs. 2
(18.2) | | | NR | 10 patients who received tranylcypromine, side
effects occurred in eight patients, including ataxia (n = 2),
dizziness (n = 3), weakness, trembling 5 due to Aes (n
= 2), high blood pressure, headaches, dry mouth, cold/hot
flushes and impotence. In the 13 patients who received lamotrigine, nine
patients reported side effects, including pain in attachment of muscles,
transpiration, itching and irritated eyes, tremor (n = 2),
diarrhoea, tiredness, obsessions and itchy rash over body | 8 withdrawals in 1st 10 weeks 5 due to
Aes | Very small, no AE assessment reported and then Aes
reported in way that cannot be compared. |
Schaffer, 2006 Canada | RCT, double blind | Inclusion- met DSM-IV criteria for bipolar
disorder (BD) type I or II, with a current major depressive episode;
outpatients, age 18–65, who spoke fluent English, and had a
baseline 17-item Hamilton Depression Rating Scale (HAM-D) score of
≥16. Patients must have been treated with a mood stabilizer for
at least the past 4 weeks Exclusion - a current hypomanic, manic, or mixed
episode, score of ≥12, current psychotic symptoms, substance
abuse/dependence during the past 3 months, current antidepressant use,
discontinuation of any mood stabilizer, antidepressant, or antipsychotic
medication within less than 5 half-lives, past treatment with lamotrigine or
citalopram in combination with current mood stabilizer(s), unstable medical
condition, history of Stevens– Johnson syndrome,
lamotrigine- induced rash, or pregnancy. | lamotrigine 81.3–100
mg/day citalopram 21 mg/day as add-on treatment
12
weeks | None | treated with a mood stabilizer for at least the
past 4 weeks (confirmed by clinical recordswhen available), including one or
more of lithium (baseline serum level ≥0.6 mmol/L), divalproex
sodium (baseline serum level ≥50 μg/mL), or
carbamazepine (baseline serum level ≥4.0
μg/mL). | HAM-D, MADRS, YMRS, CGI
Assessed at
baseline, and weeks 1, 2, 4, 6, 8, 10 and 12 | 41 years 85% female Ethnicity
NR | 65% single, 40% living
alone 70% completed at least some post-secondary
education 60% unemployed, and 65% on
disability | NR/NR/NR/20 | 8/0/19 | Change in total MADRS score (citalopram
−14.2, vs. Lamotrigine − 13.3 P =
NS | Respose and remission shown in graphs but There
were numerical, but no significant differences between treatment groups in
favor of citalopram. | | | NR | NR except for withdrawal due to. | 8 withdrawals 4 due to AEs | small n |
Suppes, 2008 USA (Texas) | RCT, open label | Inclusion - Outpatients between the ages of
18–65; clear history of BDII confirmed by SCID-research version
interview and met criteria for DSM-IV bipolar II disorder (hypomanic episode
duration of 4 or more days), current episode depressed, and currently had a
score of 18 or greater on the Hamilton Rating Scale for Depression-17 item
(Ham-D17) or ontgomery–Asberg Depression Rating Scale (MADRS).
Exclusion - history of clinically relevant intolerance or nonresponse to
lithium (Li) or lamotrigine (LTG), an unstable medical illness, psychotic
symptoms, active suicidal ideation or intent, or substance abuse or
dependency within the last month. Women who were pregnant, planning to
conceive, or breastfeeding | Lamotrigine (LTG) 900 mg/day Lithium (Li) 200
mg/day | Tapered off of former meds but no set time. | Short-term use of limited benzodiazepines/hypno
tics for a maximum of 5 consecutive days, on no more than one occasion over
the course | HAM-D, MADRS, YMRS, CGI-BP, GAF | 36 years 63%
female 77% white | | NR/NR/NR/102 | 50/18/90 | LTG avs. Li Mean Ham-D 17 (SE)
8.00±1.28 vs. 6.97±1.33 | Met response criteria (50% or greater
reduction on Ham-D-17) LTG 67.5% vs. Li 55.1%
remission rates without switch; LTG 75.6% vs. Li
59.2% | Mean baseline YMRS scores (SD) LTG
6.84±4.34 and Li 6.71±3.82 (p=0.63) mean
endpoint scores (SE) LTG 2.92±0.93 and Li 3.00±0.97
P=0.68
Mean baseline CGI-BP severity scores LTG
4.5±0.7 and Li 4.6±0.6 mean endpoint scores(SE) LTG
1.9±0.31 and Li 2.2±0.32 p=0.61 | | NR- vital signs (weight, blood pressure, and
pulse) were measured and a side effect assessment completed (but method not
reported) | Lamotrigine vs. Lithium Cognitive slowing
7.3% vs. 26.5% Decreased sexual interest
2.4% vs. 16.3% Dizziness/lightheadedness
7.3% vs. 30.6% Drowsiness/panic
9.8% vs. 30.6% Dry mouth 19.5%
vs. 53.1% Feeling dull 2.4% vs.
18.4% Impaired memory .0% vs.
20.4% Increase thirst 7.3% vs.
49.0% Increased appetite 4.9% vs.
28.6% Increased urinary frequency 2.4% vs.
32.7% Increased weight 4.9% vs.
22.5% Nausea/vomiting 24.4% vs.
46.9% Ringing in ears .0% vs.
12.2% Tremor 9.8% vs.
40.8% Upset stomach 19.5% vs.
42.9% Word finding 4.9% vs.
24.5% | 50 withdrawals 18 due to Aes | |
Placebo-Controlled Trials
|
Bowden, 2006 U.S | DB, RCT, muticenter Inpatient setting | Male and female patients, aged 18- 65y;
hospitalized for acute excaberation; current DSM-IV-TR primary diagnosis of
bipolar I disorder, manic or mixed type, confirmed by the Structured
Clincial Interview; MRS ≥ 18 with > 4 item scores over
1 | Divalproex ER qd (initial 25 mg/kg rounded to
nearest 500 mg, increased by additional 500 mg on day 3; dose adjustments on
days 7, 12, and 17 for emerging AEs, mean dose at end of study 3057 mg/d)
vs. Placebo qd for 21d | 3d minimum screening/wash-out period | No adjunctive psychotropic medciations allowed;
only lorazepam allowed during washout and treatment periods (maximum dose
was 2 mg, during screening 6 mg/d, 1–7d was 4 mg/d, and
8–10d was 2 mg/d); lorazepam not allowed within 8h before
efficacy assessment | SADS-C (including the MRS and Depressive Syndrome
Scale) and the GAS assessed on days 1, 5, 10, 15 and 21; the DSS only on the
first and last day | Divalproex vs. Placebo Age: 37.0(10.71) vs.
38.1(10.28); p=0.322 Male
(%):113(60%) vs. 96(54%);
p=0.245 Ethnicity: White: 135(72%) vs.
135(76%); p=0.403 | Divalproex vs. Placebo Weight (kg):
87.4(22.23) vs. 87.1(21.39); p=0.888
Manic
episdoe: 107(57%) vs 88(50%) Mixed episode:
80(43%) vs. 79(45%);
p=0.752 Psychotic features: 36(19%) vs.
39(22%); p=0.520
< 6 manic
episodes: 68(37%) vs. 79(45%);
p=0.032
< 6 mixed episodes:
112(61%) vs. 105(60%);
p=0.593
< 6 depressive episodes:
114(62%) vs. 104(60%);
p=0.820
Rapid cycling: 9(5%) vs.
13(7%); p=0.381
Age, first manic
episode (y): 22.6(9.8) vs. 23.8(10.8) p=0.564 Age, first
mixed episode (y): 25.4(9.7) vs. 25.7(10.8); p=0.912 Age,
first depressive episode (y): 20.9(9.1) vs. 22.2(9.7);
p=0.381
< 6 bipolar hospitalizations:
109(58%) vs. 117(66%);;
p=0.553 Age, first hospitalization (y): 26.6(10.2) vs.
28.3(11.7); p=0.284
Suicide attempts:
105(56%) vs. 95(54%); p=0.603 | Number screened not reported/number eligible not
reported/377 enrolled/377 randomized | 13 withdrawn/lost to follow-up not reported/364
analyzed (ITT population) | Divalproex vs. Placebo MRS, mean change:
−11.5(10.9) vs. − 9.0(10.9); p=0.013;
p=0.007 for treatment differences MSS, mean change:
−6.7(6.0) vs. − 5.3(6.0);
p=0.009 BIS, mean change: −4.5(5.1) vs.
− 3.4(5.1); p=0.019
Items of MRS,
mean change: More energetic: −1.3 vs. −1.0;
p<0.05 Elevated mood: −1.2 vs. 1.0 Less
need for sleep: −1.7 vs. −1.2;
p<0.01 Increased activity: −1.3 vs.
−1.0; p<0.05 Generalized motor hyperactivity:
− 1.2 vs. −0.9; p<0.05 Pressured
speech: −1.1 vs. −0.9 Grandiosity:
−0.9 vs. −0.8 Overt anger: −0.6
vs. −0.5 Poor judgement: −0.8 vs.
−0.5 Racing thoughts: −0.8 vs.
−0.5; p<0.001 Lack of insight: −0.2
vs. −0.3 | Divalproex vs. Placebo Resonders
(≥ 50% improvement MRS): 48% vs.
34%; p=0.012
Remission:
48% vs. 35%;
p=0.015
Effectiveness: 38% vs.
26%; p=0.032
Discharged from
hospital: 17% vs. 20% | MRS, mean change, treatment interaction signficant
(p=0.006)
Divalproex vs. Placebo Higher
baseline MSS scores (>13): −15.0 vs.
−9.6 Lower baseline MSS scores (< 13):
−7.5 vs. −7.7
Treatment difference
not a function of age, gnder, race, episode type, presence of psychotic
features, rapid cycling, previous episodes, drug abuse or baseline DSS | Divalproex vs. Placebo Rescue medicaton:
74% vs. 79%
Dose of rescue medication
higher in placebo group than divalproex group on day 1; similar on all other
days | AEs monitored throughout study; hematological,
blood chemistry, vital signs, physical examination and urinalysis performed
periodically | Divalproex vs. Placebo AEs:
162(84%) vs. 134(72%);
p=0.006 Somnolence: 64(33%) vs.
26(14%); p<0.001 Nausea: 53(28%) vs.
28(15%); p=0.004 Dyspepsia:
49(26%) vs. 18(10%); p<0.001 Headache:
40(21%) vs. 40(22%);
p=0.9 Dizziness: 39(19%) vs.
15(8%); p=0.003 Vomiting: 35(18%)
vs. 12(6%); p=0.001 Diarrhea:
28(15%) vs. 18(10%); p=0.16 Pain:
23(12%) vs. 16(9%);
p=0.314 Abdominal pain: 19(10%) vs.
8(4%); p=0.045 Pharyngitis:
19(10%) vs. 8(4%);
p=0.045
Mean changes in laboratory parameters
(p<0.05): RBC: −0.01(0.28) vs.
0.05(0.28) Platelets: −58.8(59.9) vs.
−1.2(49.6) Moncytes: 1.88(3.03) vs.
0.29(2.37) Basophils: −0.05(0.23) vs.
0.02(0.25) Protein (g/dL): −0.27(0.51) vs.
0.03(0.52) Albumim (g/dL): −0.19(0.31) vs.
0.02(0.28) Bilirubin (mg/dL): −0.09(0.21) vs.
0.01(0.26) Alkaline phosphatase (IU/L): −12.15(13.07) vs.
−1.58(14.91) Aspartate aminotransferase (IU/L):
−3.93(11.42) vs. −0.19(10.32) Alanine
aminotransferase (IU/L): −6.41(18.13) vs.
1.92(17.76) Sodium (mEq/L): 0.64(2.63) vs. 0.03(2.85) Calcium
(mEq/L): −0.2(0.43) vs. 0.07(0.43)
Weight gain
(kg): 1.8(3.43) vs. 0.5(2.89); p<0.001 BMI (kg/m2): 0.61(1.14)
vs. 0.14(1.06); p<0.001 Fasting glucose (mg/dL):
−1.24(35.49) vs. 1.71(26.1);
p=0.371 Cholesterol (mg/dL): −13.47(30.93)
vs. −2.46(32.76); p=0.001 HDL (mg/dL):
−6(9.84) vs. −2.79(10.56);
p=0.001 LDL (mg/dL): −7.89(28.92) vs.
−0.07(29.9); p=0.003 Triglycerides (mg/dL):
1.53(90.73) vs. −4.34(98.66); p=0.556 | Divalproex vs. Placebo Total withdrawals:
80/192(42%) vs. 89/185(48%) Withdrawals due
to AEs: 19/192(10%) vs. 6/185(3%);
p=0.003 | Those discontinued due to AEs had hgihe mean serum
valproate concentrations than patients who did not (p=0.012);
those who discontinued due to GI AEs also had higher mean serum valproate
concentrations (p=0.002)
Two cases of
pancreatitis reported (one during study, one after); one death in divalproex
group (not related to study drug) |
Chengappa, 2006 U.S. | Multicenter, randomized, double-blind, placebo-
controlled, parallel-group study Outpatient setting | Adults aged 18–70y; bipolar I disorder
(defined by the DSM-IV carier and supported by the Structured Clinical
Interview for the DSM-IV Axis I Disorders and the YMRS score ≥
18); received either lithium or valproate for > 6w; erm level of mood
stabilizers between 0.5–1.2 mEq/L for lithium and
45– 100 mg/L for valproate | Topiramate 25 mg/d (titrated weekly −
50, 75, 100, 150, 200, 300, and 400 mg/d; adjusted for any emerging AEs;
mean dose 254.7 mg/d) vs. Placebo for 12w | Not reported | Lithium or divalproex sodium; permitted to
continue taking a stable dose of an oral antipsychotic agent; use of
short-acting benzodiazepine (lorazepam) for sleep or agitation permitted
only during first 4w of titration period
Topiramate vs.
Placebo Valportate: 91(66.9%) vs.
78(54.9%) Lithium: 45(33.1%) vs.
64(43.0%) Lorazepam: 15(11.0%) vs.
23(16.2%) Alprazolam: 2(1.5%) vs.
2(1.4%) Clonazepam: 2(1.5%) vs.
0(0%) Escitalopram: 2(1.5%) vs.
0(0%) Sertraline: 2(1.5%) vs.
1(0.7%) Bupropion: 1(0.7%) vs.
3(2.1%) Paroxetine: 1(0.7%) vs 2(1
4%) | YMRS, Clinical Global Impressions-Severity of
Illness scale (CGI-S), Brief Psychiatric Rating Scale (BPRS),
Montgomery-Asberg Depression Rating Scale (MADRS) and the Global Assessment
scale assessed at each visit | Topiramate vs. Placebo Age: 41.0(12.2) vs.
39.0(11.9) Male (%):58(40.6%) vs.
67(46.5%) Ethnicity: White: 119(83.2%) vs.
122(84.7%) Black: 14(9.8%) vs.
14(9.7%) Hispanic: 5(3.5%) vs.
6(4.2%) Asian: 2(1.4%) vs.
1(0.7%) Other: 3(2.1%) vs.
1(0.7%) | Topiramate vs. Placebo BMI (kg/m2): 31.0(7.8) vs.
30.4(7.3)
Rapid cycling: 39(27.3%) vs.
43(29.9%) Mania: 105(73.4%) vs.
102(70.8%) Mixed: 30(21.0%) vs.
35(24.3%) Missing: 0(0%) vs.
7(4.9%)
Hospitalizations: 2.7(5.8) vs.
3.3(13.1)
Psychotic episodes: 42(29.4%) vs.
36(25.0%) Number of psychotic episodes: 7.3(20.0) vs.
7.8(17.1) | 424 screened/number eligible not reported/287
enrolled/287 randomized | 110 withdrew/31 lost to follow-up/278 analyzed
(ITT population) | Topiramate vs. Placebo YMRS, mean change:
−10.1(8.7) vs. 9.6(8.2);
p=0.797
CGI-S, mean change: −0.9(1.1)
vs. − 0.9(1.1); p=0.698
BPRS, mean
change: −3.3(9.6) vs. − 4.8(9.0);
p=0.052
MADRS, mean change: 0.6(8.8) vs.
− 1.1(9.0); p=0.057
GAS, mean change:
7.2(9.9) vs. 7.1(11.5); p=0.838 | Topiramate vs. Placebo Response rate,
> 50% reduction in YMRS: 39.0% vs.
38.0%; p=0.914 | | | Reports of AEs and vital signs at every visit;
clinical laboratory tests and serum pregnancy tests every 4w and at the end
of treatment; serum levels of lithium, valproate and topiramate at baseline,
end of dose titration and end of study | Topiramate vs. Placebo AEs:
122(85.3%) vs. 120(83.9%) Headache:
34(23.8%) vs. 37(25.9%) Paresthesia:
33(23.1%) vs. 5(3.5%); p<0.05 Upper
respiratory tract infection: 25(17.5%) vs.
16(11.2%) Diarrhea: 24(16.8%) vs.
12(8.4%); p<0.05 Nausea: 22(15.4%) vs.
17(11.9%) Somnolence: 22(15.4%) vs.
23(16.1%) Anorexia: 19(13.3%) vs.
8(5.6%); p<0.05 Insomnia: 17(11.9%)
vs. 16(11.2%) Memory difficulty: 16(11.2%)
vs. 10(7.0%) Dizziness: 15(10.5%) vs.
15(10.5%) Abnormal vision: 12(8.4%) vs.
7(4.9%); NSD between groups Suicidality:
1(0.7%) vs. 2(1.4%) Death: 0(0%)
vs. 0(0%)
Clinical laboratory visits and vital
signs were within normal range
Body weight, mean change (kg):
−2.5(3.4) vs. 0.2(3.0); p<0.001 BMI, mean change
(kg/m2): −0.84(1.2) vs. 0.07(1.1); p<0.001 | Topiramate vs. Placebo Total withdrawals:
57/143(39.9%) vs. 53/144(36.8%) Withdrawals
due to AEs: 20/143(13.9%) vs. 10/144(6.9%) | Topiramate vs. Placebo Follow-up (d):
70.8(31.6) vs. 74.7(30.0) out of total 91 possible
days
Adjunctive treatment with topiramate not associated with
worsening of symptoms; similar poritons of patients in each group
experienced worsening symptoms; NSD between groups |
Frankenburg, 2002 U.S. | Placebo-controlled, double-blind
study Outpatient setting | Women aged 18–40y; distubred by mood
changes, distrustfulness, impulsivity and stormy relationships; DSM-IV
criteria for borderline personality disorder using the borderline module of
the Diagnostic Interview for DSM-IV Personality Disorders; Structured
Clinical Interview for DSM-IV Axis I Disorder and Revised Diagnostic
Interview for Borderlines and DSM-IV bipolar I disorder criteria (not
curretnly in depressive or hypomanic episodes)
With no
previous treatments with divalproex sodium, medical illness, alcohol and
drug abuse; not pregnant, breastfeeding, planning to become pregnant or
having unprotected sex | Divalproex sodium 250 mg bid (mean dose, 850 mg/d)
vs. Placebo for 6m | Not reported | No other psychotropic drug allowed during the
study | Two self report measures: Symptom Checklist 90
(SCL-90) and the McLean version of the modified Overt Aggression Scale
(MOAS) Checklist measured at each visit | Divalproex vs. Placebo Age: 27.3(7.4) vs.
26.4(7.3); NSD Male (%): 0(0%) vs.
0(0%) Ethnicity: White: 15(75%) vs.
5(50%); NSD Black: 3(10%) Hispanic:
4(13.3%) Biracial: 2(6.7%) | Divalproex vs. Placebo GAF score: 51.6(6.5)
vs. 50.2(7.0); NSD
Individual therapy: 12(60%)
vs. 7(70%); NSD Medication: 9(45%) vs.
3(30%); NSD Hospitalized: 1(5%) vs.
2(20%); NSD | Number screened not reported / number eligible not
reported / 30 enrolled / 30 randomized | 19 withdrew / 12 lost to followup / 13 analyzed
(20 analyzed at 8w) | Divalproex vs. Placebo SCL-90
interpersonal: −31.7% vs.
−14.8%; p=0.0408
SCL-90
anger/hostility: −29.6% vs. −11.0;
p=0.0117
SCL-90 depression:
−21.3% vs. −25.4%;
NSD
MOAS total: −42.1% vs.
−13.4%; p=0.0278 | Weight gain, mean change (lbs): 2.6(5.6) vs.
0.3(4.0); p=0.1175
Weight gain (%):
1.9(3.9) vs. 0.12(3.1); p=0.1185 | | | Monitored | Divalproex vs. Placebo Major depressive
episode: 0(0%) vs. 2(20%) Menstrual changes:
1(5%) vs. 1(10%); NSD Tremors/diarrhea:
1(5%) vs. 0(0%) Hair loss: 0(0%)
vs. 1(10%) Increase hepatic transaminases:
1(5%) vs. 0(0%) Thrombocytopenia:
0(0%) vs. 0(0%) | Divalproex vs. Placebo Total withdrawals:
13/20(65%) vs. 6/10(60%) Withdrawals due to
AEs: 1/20(5%) vs. 3/10(30%) | 12h trough levels were done at 1w, 1m and then
every 2m for dose adjusting
Due to high level of attrition at
6m, analyses repeated using only data collected up to 8w |
Pope, 1991 U.S | Single center, placebo-controlled,
double-blind Inpatient setting | Aged 18–65y, meeting DSM-III-R
criteria for bipolar disorder, manic phase; failure to respond adequately to
a trial of lithium or intolerance of lithium side effects; females must be
postmenopausal or surgically sterilized | Divalproex sodium 250mg tid vs. Placebo for
7–21 days | Not reported | Treatment with all psychotropic medication
discontinued; only lorazepam 1 mg qid allowed to treat agitation or insomina
up to day 9 | YMRS, Global Assessment Scale (GAS), Brief
Psychiatric Rating Scale, Augmented (BPRS-A) on days 7, 14, 21 | Valproate vs. Placebo Age: 39.7(11.8) vs.
34.6(14.7) Male (%):13(76.5%) vs.
13(68.4%) Ethnicity: Not reported | Valproate vs. Placebo Duration of illness
(y): 12.2(10.9) vs. 11.2(9.7)
MRS: 28.2(5.8) vs.
28.6(6.9) GAS: 30.0(5.9) vs. 31.6(5.5) | Number screened not reported / number eligible not
reported / 43 enrolled / 43 randomized | 35 withdrew / 0 lost to follow-up / 36 analyzed
(ITT efficacy analysis of all 43 patients) | Valproate vs. Placebo
YMRS, median
improvement: 54% vs. 5.0%;
p=0.003
GAS, improvement (points): 20 vs 0;
p=0.002
BPRS-A, median improvement: 17 vs 3;
p=0.001
Response rate, > 50%
improvement: 9(53%) vs. NR | On four of the 18 standard BPRS subscales
(conceptual disorganization, tension, hostility, excitement), valproate
improved significantly more than placebo (p<0.005); as well as
increased motor activity (p=0.006); no subscale showed
significant change in favor of placebo | Analysis of covariance of patients score at
termination, valproate patients improved significantly more than placebo on
the YMRS (p=0.005), GAS (p=0.001) and the BPRS-A
(p=0.001) | ITT analysis showed that valproate patients
improved more significantly on the YMRS (p=0.013), GAS
(p=0.004) and total BPRS-A (p=0.002) than the
placebo patients
Subgroup analysis of patients receiving
lorazepam on valproate (n=13) showed a significantly greater
improvement on the YMRS (p=0.016), GAS (p=0.008) and
the BPRS-A (p=0.002) than placebo patients
(n=12) | Blood chemistry studies, hematologic studies and
urinalyses repeated at days 7, 14 and 21 | Valproate (n=20) vs. Placebo
(n=23) GI discomfort with vomiting: 5(25%)
vs. 5(22%) GI discomfort without vomiting:
1(5%) vs. 2(8.7%) Headache:
4(20%) vs. 6(26%) Sedation/fatigue:
4(20%) vs. 1(4.3%) Constipation:
0(0%) vs. 3(13%) Swelling/pain:
1(5%) vs. 2(8.7%) Ataxia: 2(10%)
vs. 0(0%) Dysuria: 0(0%) vs.
2(8.7%) Palpitations: 1(5%) vs.
1(4.3%) Diplopia: 1(5%) vs.
1(4.3%) Tightness in chest: 1(5%) vs.
0(0%) Dry eyes: 1(5%) vs.
0(0%) Sinus pressure: 1(5%) vs.
0(0%) Dysarthria: 1(5%) vs.
0(0%) Depression: 1(5%) vs.
0(0%) Diarrhea: 1(5%) vs.
0(0%) Anorexia: 1(5%) vs.
0(0%) Agitation: 1(5%) vs.
0(0%) Bruising: 0(0%) vs.
1(4.3%) Lump in throat: 0(0%) vs.
1(4.3%) Panic attacks: 0(0%) vs.
1(4.3%) | Valproate vs. Placebo Total withdrawals:
13/17(76.5%) vs. 15/19(78.9%) Withdrawals due
to AEs: 2/17(12%) vs. 1/19(5.3%);
[5/43(11.6%) included patients who withdrew before
7d)
Withdrawals defined as those who did not complete 21
days | Presence of unblinded investigator adjusted dosage
to achieve desired serum concentrations; also performed sham adjustments; he
was informed if any patient complained of an AE; this investigator broke the
blinding by informing ward physician and staff of allocation when a patient
withdrew; investigator performing rating remained
blinded
Valproate vs. Placebo Follow-up (d): 13.6(4.9) vs.
12.4(5.0) --36/43 patients completed at least 7 days, only those
analyzed
Individual patients characteristics listed in table
1; figures 1 and 2 describe change in the YMRS and
GAS; valproate showed greatest improvement |
Vieta, 2006 Spain | Multicenter, double-blnd, randomized, placebo-
controlled, parallel-group study | Aged 18–75y with diagnosis of bipolar
I or II disorder (according to the DSM-IV criteria) treated with any
standard mood stabilizer; > 2 bipolar episodes during last year; CGI
scale for Bipolar Illness, Modified score ≥ 4 with last episode
occurring within past 6m; euthymic at randomization (Hamilton Rating Scale
for Depression ≤ 8 and YMRS ≤ 4) | Gabapentin 1200 mg/d (with emerging symptoms,
increased to 2400 mg/d; with AEs, reduced to 900 mg/d) vs. Placebo for
12m | Not reported | Lithium, valproate, carbamazepine or any
combination; no treatment with antipsychotics or antidepressants
allowed | Clinical Global Impressions scale for Bipolar
Illness, Modified (CGI-BP-M); 7-point ranging scale (1-not ill, 7-extremely
ill); other assessments included the YMRS, Hamilton Rating Scale for
Depression (HAM-D), Hamilton Rating Scale for Anxiety (HAM-A), Pittsburgh
Sleep Quality Index (PSQI) at each visit | Gabapentin vs. Placebo Age: 46.2(14.3) vs.
47.6(15.8) Male (%):3(23.1%) vs.
4(33.3%) Ethnicity: Not reported | Gabapentin vs. Placebo Weight (kg):
74.6(13.8) vs. 63.8(12.1)
Seasonal pattern:
3(25.0%) vs. 2(16.7%) Rapid cycling:
5(38.5%) vs. 6(50.0%)
Bipolar II:
1(7.7%) vs. 5(41.7%) Diagnosis (y):
20.9(11.5) vs. 16.5(10.5)
Episodes, total: 33.8(25.1) vs.
17.8(18.7) Manic: 6.8(8.3) vs. 4.1(6.3) Hypomanic: 6.(7.9) vs.
5.1(7.6) Depressive: 19.3(19.0) vs. 8.3(7.9) Mixed: 0.8(1.6)
vs. 0.4(0.9)
Hospitalizations: 4.1(5.4) vs. 2.4(2.3) | Number screened not reported/number eligible not
reported / 25 enrolled / 25 randomized | 12 withdrew / 0 lost to follow-up /25
analyzed | Gabapentin vs. Placebo
CGI-BP-M,
mean change: −2.1 vs. − 0.6;
p=0.0046
YMRS, mean change: 3.1 vs.
−0.6; p=0.2038
HAM-D, mean change:
1.3 vs. 2.5; p=0.6753
HAM-A: −0.3 vs.
−0.9; p=0.8443 | Gabapentin vs. Placebo PSQI, mean change:
−1.3 vs. 0.2; p=0.3362 PSQI-1, mean change:
−0.1 vs. 0; p=0.7649 PSQI-2, mean change: 0
vs. 0.4; p=0.3117 PSQI-3, mean change: 0.3 vs. 0.2;
p=0.7888 PSQI-4, mean change: 0.2 vs. 0.3;
p=0.5518 PSQI-5, mean change: 0 vs. 0;
p=0.9521 PSQI-6: −1.1 vs. −0.6;
p=0.0267 PSQI-7: −0.3 vs. −0.2;
p=0.7842 | Gabapentin vs. Placebo Time from randonmization to
first new episode showed NSD between groups (p=0.6658); HR
1.344 | | Side effects were systematically collected | Gabapentin vs. Placebo AEs reported:
10(77%) vs. 7(58%) Constipation:
4(31%) vs. NR Headache: 3(23%) vs.
NR Nausea: 3(23%) vs. NR Dizziness:
2(15%) vs. NR Insomnia: 2(15%) vs.
NR Tremor: 2(15%) vs. NR | Gabapentin vs. Placebo Total withdrawals:
7/13(54%) vs. 6/12(50%) Withdrawals due to
AEs: 1/13(7.7%) vs. 1/12(8.3%) | Prophylactic study |
GSK (SCA100223) U.S. | Double-blind, multicenter, placebo-controlled,
fixed-dose, parallel study Outpatient setting | Male or non-pregnant females >18 years with
a diagnosis of bipolar II disorder who were currently experiencing a major
depressive episode of at least 8w duration; HAMD−17 > 18;
HAMD- item1 or HADM-item7 > 3; no suicidal activity | Lamotrigine 200 mg/d vs. Placebo for 8w | Not reported | Not reported | Montogomery-Asberg Depression Rating Scale
(MADRS); Clinical Global Impressions of Improvement (CGI-I); Hamilton
Depression Rating Scales; Beck Melancholia Scale; Clinical Global
Impressions of Severity (CGI-S) and Mania Rating Scale from the Schedule of
Affective Disorders and Schizophrenia (MRS) and the Treatment Satisfaction
Question at 8w | Lamotrigine vs. Placebo Age: 38.1(11.5) vs.
36.5(1.9) Male (%): 39(35.8%) vs.
40(36.7%) Ethnicity: White: 70(64%) vs.
82(75%) | | Number screened not reported / number eligible not
reported / 221 enrolled / 221 randomized | 66 withdrew / lost to follow-up not reported / 214
analyzed | Lamotrigine vs. Placebo
MADRS, mean
change: −13.4(1.0) vs.
−12.0(1.0)
HAMD−17, mean change:
− 11.1(0.8) vs. −9.4(0.8)
HAMD-31,
mean change: −16.0(1.1) vs.
−13.8(1.2)
HAMD-item 1, mean change:
−1.4(0.1) vs. −1.3(0.1) | Lamotrigine vs. Placebo
BMS, mean
change: −6.8(0.5) vs.
−5.4(0.5)
CGI-S, mean change:
−1.4(0.1) vs. −1.3(0.1)
CGI-I, mean
change: 2.5(0.1) vs. 2.8(0.1)
MRS-16, mean change:
− 0.4(0.3) vs. 0.1(0.3) | Lamotrigine vs. Placebo
Responders,
MADRS: 59(54.1%) vs.
48(45.7%)
Responders, HAMD−17:
56(51.4%) vs. 42(40%)
Responders,
BMS: 62(56.9%) vs.
47(44.8%)
Responders: CGI-I:
66(60.6%) vs. 47(44.8%) | Lamotrigine vs. Placebo
Remission,
MADRS: 45(41.3%) vs.
36(34.3%)
Remission, HAMD−17:
28(25.7%) vs. 29(27.6%) | Treatment emergent AES reported at each study
treatment visit and at follow-up visit | Lamotrigine vs. Placebo Any AEs:
88(81%) vs. 85(78%) Headache:
30(28%) vs. 39(36%) Dry mouth:
10(9%) vs. 7(6%) Insomnia: 9(8%)
vs. 7(6%) Diarrhea: 8(7%) vs.
18(17%) Nasopharyngitis: 8(7%) vs.
9(8%) Nausea: 8(7%) vs.
15(14%) Dizziness: 7(6%) vs.
8(7%) Rash: 7(6%) vs.
7(6%) Cough: 6(6%) vs.
5(5%) Sedation: 6(6%) vs.
2(2%) Somnolence: 6(6%) vs.
5(5%) Fatigue: 3(3%) vs.
7(6%) Vomiting: 2(2%) vs.
6(6%)
Non-factal serious AEs: 0(0%)
vs. 5(5%) Pneumonia: 0(0%) vs.
1(1%) Suicidal ideation: 0(0%) vs.
2(2%) Suicide attempt: 0(0%) vs.
1(1%) Agitation: 0(0%) vs.
1(1%) Cellulitis: 0(0%) vs.
1(1%) Fatal AEs: 0(0%) vs.
0(0%) | Lamotrigine vs. Placebo Total withdrawals:
30/111(27%) vs.
36/110(33%)
Withdrawals due to AEs:
4/111(4%) vs. 5/110(5%) | Serious AES not related to study drug |
GSK (SCA30924) U.S. | Double-blind, multicenter, placebo-controlled,
fixed- dose, parallel study Outpatient setting | Male or non-pregnant females > 18 years
with a diagnosis of bipolar II disorder who were currently experiencing a
major depressive episode of at least 8w duration; HAMD-17 > 18; HAMD-
item1 or HADM-item7 > 3; no suicidal activity | Lamotrigine 200 mg/d vs. Placebo for 8w | Not reported | Not reported | Montogomery-Asberg Depression Rating Scale
(MADRS); Clinical Global Impressions of Improvement (CGI-I);
Hamilton Depression Rating Scales; Beck Melancholia Scale;
Clinical Global Impressions of Severity (CGI-S) and
Mania Rating Scale from the Schedule of Affective Disorders
and Schizophrenia (MRS) and the Treatment Satisfaction Question at 8w | Lamotrigine vs. Placebo Age: 40.5(12.5) vs.
38.2(12.1) Male (%): 58(45.7%) vs.
56(45.9%) Ethnicity: White: 94(74%) vs.
84(69%) Black: 18(14%) vs.
26(21%) | | Number screeneed not reported / number eligible
not reported / 259 enrolled / 259 randomized | 107 withdrew / lost to follow-up not reported /
243 analyzed | Lamotrigine vs. Placebo
MADRS, mean
change: −12.6(1.0) vs.
−11.7(1.0)
HAMD-17, mean change:
−9.8(0.7) vs. −9.3(0.7)
HAMD-31, mean
change: − 15.0(1.1) vs.
−13.7(1.1)
HAMD-item1, mean change: −
1.3(0.1) vs. −1.2(0.1) | Lamotrigine vs. Placebo
BMS, mean
change: 13.3(2.0) vs. 13.4(2.2)
CGI-S, mean change:
−5.5(0.4) vs. −5.2(0.4)
CGI-I, mean
change: 2.8(0.1) vs. 2.8(0.1)
MRS-16, mean change:
− 0.1(0.3) vs. −0.1(0.3) | Lamotrigine vs. Placebo
Responders,
MADRS: 56(45.5%) vs.
48(40.0%)
Responders, HAMD-17:
51(41.5%) vs. 39(32.5%)
Responders,
BMS: 52(42.3%) vs.
48(40.0%)
Resonders, CGI-I: 59(48.0%)
vs. 47(39.2%) | Lamotrigine vs. Placebo
Remission,
MADRS: 33(26.8%) vs.
36(30.0%)
Remission, HAMD-17:
18(14.6%) vs. 29(24.2%) | Treatment emergent AEs were reported at each study
treatment visit | Lamotrigine vs. Placebo Any AEs:
95(75%) vs. 87(71%) Headache:
26(20%) vs. 26(21%) Diarrhea:
17(13%) vs. 7(6%) Nausea: 14(11%)
vs. 12(10%) Dry mouth: 13(10%) vs.
8(7%) Somnolence: 10(8%) vs.
4(3%) Back pain: 7(6%) vs.
7(6%) Dizziness: 8(6%) vs.
2(2%) Rash: 8(6%) vs.
2(2%) Insomnia: 6(5%) vs.
9(7%) Nasopharyngitis: 4(3%) vs.
10(8%)
Non-fatal serious AEs: 4(3%)
vs. 6(5%) Mania episode: 1(1%) vs.
1(1%) Mixed manic depressive episode: 0(0%)
vs. 1(1%) Suicidal ideation: 1(1%) vs.
2(2%) Agitation: 1(1%) vs.
1(1%) Acute stress disorder: 0(0%) vs.
1(1%) Depression: 0(0%) vs.
1(1%) Limb injury: 0(0%) vs.
1(1%) Muscle injury: 0(0%) vs.
1(1%) Vomiting: 0(0%) vs.
1(1%) Abdominal pain: 1(1%) vs.
0(0%) Hepatic encephalopathy: 1(1%) vs.
0(0%) Depressive symptom: 1(1%) vs.
0(0%) Fatal AEs: 0(0%) vs.
0(0%) | Lamotrigine vs. Placebo Total withdrawals:
52/131(40%) vs. 55/128(43%) Withdrawals due
to AEs: 13/131(10%) vs. 9/128(7%) | Serious AES not related to study drug |
GSK (SCAA2010) U.S. | Multicenter, double-blind, placebo-controlled,
flexible dose trial Outpatient setting | Subjects > 18 years; diagnosis of bipolar I
or II disorder; current episode depressed (defined by the DSM-IV, based on
modified Structured Clinical Interview for DSM-IV); currently experiencing a
major depressive episode; one major depressive, manic or mixed episode in
past 10y; or two hypomanic episodes (for bipolar II) in past 10y; duration
of current episode 2–52w; HAMD-17 > 18 | Lamotrigine flexible dose (100–400
mg/d, target dose 360 mg/d) vs. Placebo for 10w | Not reported | Not reported | 17-item Hamilton Depression Rating Scale
(HAMD-17); HAMD-31; HAMD-item1; Montgomery-Asberg Depression Rating Scale
(MADRS); Mania Rating Scale (MRS); Schedue for Affective Disorders and
Schizophrenia-Change Version (SADS-C); and the Clinical Global Improvession
of Improvement (CGI-I) and Severity (CGI-S) at all visits | Lamotrigine vs. Placebo Age: 40.5(11.3) vs/
40.9(11.2) Male (%):37(35.9%) vs.
42(40.8%) Ethnicity: White: 90(87%) vs.
89(86%) | | Number screened not reported / number eligible not
reported / 206 enrolled / 206 randomized | 69 withdrew / lost to follow-up not reported / 202
analyzed (204 for safety) | Lamotrigine vs. Placebo
HAMD-17,
mean change: 4d: −2.6(4.4) vs. −2.7(4.4);
p=0.674 8d: −4.3(5.7) vs.
−4.1(5.4); p=0.805 15d: −5.7(5.6)
vs. −6.1(5.8); p=0.607 22d:
−7.5(5.6) vs. −7.5(6.7);
p=0.775 29d: −8.3(6.6) vs.
−8.4(7.8); p=0.804 36d: −8.8(7.3)
vs. −9.0(7.4); p=0.717 43d:
−9.4(7.4) vs. −10.0(7.9);
p=0.388 50d: −10.0(7.9) vs.
−10.1(7.9); p=0.661 57d:
−9.5(8.0) vs. −10.1(7.6);
p=0.414 64d: −10.4(8.1) vs.
−10.2(7.9); p=0.932 71d:
−10.2(8.3) vs. −10.6(8.1);
p=0.710
HAMD-31, mean change: 4d:
−4.0(6.8) vs. −4.0(6.3) 8d:
−6.5(8.8) vs. −6.1(8.2) 15d:
−7.7(8.7) vs. −8.6(9.0) 22d:
−10.8(9.0) vs. −11.0(10.1) 29d:
−11.8(10.1) vs. −12.4(11.4) 36d:
−12.7(11.1) vs. −13.2(10.8) 43d: 13 7(11 4)
vs 15 1(11 9) | Lamotrigine vs. Placebo HAMD-item1, mean
change: 4d: −0.2(0.7) vs. −0.2(0.8) 8d:
−0.4(0.9) vs. −0.4(0.8) 15d:
−0.7(1.1) vs. −0.5(0.9) 22d:
−0.8(1.1) vs. −0.7(1.0) 29d:
−0.9(1.1) vs. −0.9(1.0) 36d:
−0.9(1.1) vs. −0.9(1.0) 43d:
−1.1(1.2) vs. −1.0(1.1) 50d:
−1.1(1.2) vs. −1.0(1.1) 57d:
−1.1(1.3) vs. −1.0(1.1) 64d:
−1.2(1.2) vs. −1.0(1.2) 71d:
−1.2(1.2) vs. −1.1(1.2)
SAD-C, 16
item, mean change: 4d: 0.4(3.5) vs. −0.3(3.5) 8d:
−0.4(4.3) vs. −0.7(3.5) 15d:
−0.2(4.5) vs. −0.6(2.9) 22d:
−0.7(4.7) vs. −0.8(3.7) 29d:
−0.9(4.7) vs. −1.0(4.0) 36d:
−0.5(5.3) vs. −1.3(4.1) 43d:
−0.3(5.8) vs. −1.3(4.2) 50d:
−1.0(5.2) vs. −0.9(5.6) 57d:
−0.8(5.6) vs. −1.3(4.5) 64d:
−0.9(5.6) vs. −1.3(5.1) 71d:
−0.8(5.9) vs. −1.6(4.9) | Lamotrigine vs. Placebo
MADRS, mean
change: 4d: −2.8(5.4) vs. −2.6(6.5) 8d:
−4.7(7.2) vs. −4.1(7.4) 15d:
−6.1(8.4) vs. − 6.4(7.9) 22d:
−8.1(7.9) vs. − 8.3(9.3) 29d:
−9.4(8.9) vs. − 10.0(9.3) 36d:
−9.7(9.4) vs. − 10.5(9.0) 43d:
−10.4(9.5) vs. − 11.5(9.7) 50d:
−11.7(10.8) vs. − 11.5(9.9) 57d:
−10.9(11.2) vs. − 11.6(9.9) 64d:
−12.4(11.4) vs. − 11.4(9.9) 71d:
−12.1(11.2) vs. − 12.3(12.3)
CGI-S,
mean change: 4d: −0.2(0.4) vs. −0.1(0.5) 8d:
−0.3(0.7) vs. −0.2(0.6) 15d:
−0.4(0.7) vs. − 0.4(0.8) 22d:
−0.6(0.9) vs. − 0.6(0.9) 29d: 0 8(1 0)
vs | Lamotrigine vs. Placebo
CGI-I, mean
total score: 4d: 3.7(0.6) vs. 3.8(0.7) 8d: 3.6(1.0) vs.
3.6(0.8) 15d: 3.4(1.0) vs. 3.4(0.8) 22d: 3.2(1.1) vs.
3.3(0.9) 29d: 3.1(1.2) vs. 3.1(1.0) 36d: 3.1(1.2) vs.
3.0(1.0) 43d: 2.9(1.3) vs. 2.8(1.1) 50d: 2.9(1.3) vs.
2.9(1.1) 57d: 3.0(1.4) vs. 2.8(1.1) 64d: 2.8(1.4) vs.
2.8(1.2) 71d: 2.9(1.4) vs. 2.8(1.2)
SAD-C, 11 item,
mean change: 4d: 0.6(3.2) vs. − 0.3(3.1) 8d: 0.1(3.7) vs.
− 0 5(3 0) | Subjects who received at least one dose of study
drug evaluated for safety | Lamotrigine vs. Placebo Any AEs:
97(94%) vs. 89(88%) Headache:
42(41%) vs. 37(37%) Nausea:
21(20%) vs. 31(31%) Somnolence:
18(17%) vs. 10(10%) Dizziness:
18(17%) vs. 8(8%) Rash: 17(17%)
vs. 12(12%) Infection: 17(17%) vs.
11(11%) Insomnia: 14(14%) vs.
10(10%) Pain: 12(12%) vs.
9(9%) Xerostomia: 12(12%) vs.
6(6%) Influenza: 11(11%) vs.
15(15%) Diarrhea: 9(9%) vs.
13(13%) Accidental injury: 8(8%) vs.
9(9%)
Serious AEs: 4(4%) vs.
4(4%) Attempted suicide: 1(1%) vs.
0(0%) Cancer: 1(1%) vs.
0(0%) Mania: 1(1%) vs.
1(1%) Suicide: 1(1%) vs.
0(0%) Chest pain: 0(0%) vs.
1(1%) Emotional liability: 0(0%) vs.
1(1%) Visual field defect: 0(0%) vs.
1(1%) Fatal AEs: 1(1%) vs.
0(0%) | Lamotrigine vs. Placebo Total withdrawals:
35/103(34%) vs. 34/103(33%) Withdrawals due
to AEs: 14/103(14%) vs. 8/103(8%) | |
GSK (SCA40910) U.S. | Multicenter, double-blind, placebo-controlled,
randomized, fixed-dose trial; followed by an open- label continuation
phase Outpatient setting | Males and females > 18 years in generally
good physical health; diagnosis of bipolar I disorder, current depressed
episode, as defined by DSM-IV and the Structured Clinical Interview for
DSM-IV; currnly experiencing a major depressive episode, > 2 mood
episodes in past 5y and > 1 manic or mixed episode; duration of
current depressive episode 2-52w; HAMD-17 > 18 | Lamotrigine 200 mg/d vs. Placebo (during
continuatin phase, placebo patients titrated to 200 mg/d) for 8w (21w of a
continuation phase) | Not reported; after double-blind phase, 5w blinded
transition to open label medication | Not reported | 17-item Hamilton Depression Rating Scale
(HAMD-17); HAMD-31; HAMD-item1; Montgomery-Asberg Depression Rating Scale
(MADRS); Mania Rating Scale (MRS); Schedue for Affective Disorders and
Schizophrenia-Change Version (SADS-C); and the Clinical Global Improvession
of Improvement (CGI-I) and Severity (CGI-S) and the Personal Global
Impression of Improvement (PGI-I) and the Quality of Life in Depression
Scale (QLDS) | Lamotrigine vs.
Placebo
Double-blind phase: Age:
37.6(12.6) vs. 37.3(11.5) Male (%):55(42.6%)
vs. 56(47.5%) Ethnicity: White: 37.6(12.6%)
vs. 23.7(11.5%)
Continuation
phase: Age: 38.585(12.3) Male
(%):73(47.1%) Ethnicity: White:
134(86%) | | Double-blind phase: Number
screened not reported/number eligible not reported/257 enrolled/257
randomized
Continuation phase: Number
screened not reported/number eligible not reported/161 enrolled | Double-blind phase: 85
withdrew/lost to follow-up not reported/243
analyzed
Continuation phase: 56
withdrew/lost to follow-up not reported/153 analyzed | Double-blind
phase Lamotrigine vs. Placebo MADRS, mean change:
−12.2 vs. − 11.2;
p=0.523
HAMD-17, mean change: −9.3
vs. − 8.7
HAMD-31, mean change: −
14.4(12.7) vs. −13.0(12.1)
HAMD-item1, mean
change: −1.2 vs. −1.0
CGI-S, mean
change: −1.2(1.4) vs.
−1.1(1.3)
CGI-I: 2.9 vs.
2.9
MRS-16, mean change: 0.6(6.2) vs.
−0.7(5.4)
MRS-11, mean change: 1.0(5.4) vs.
0.0(4.7) | Double-blind
phase Lamotrigine vs. Placebo QLDS, mean change:
− 10.8(10.4) vs. −7.0(10.7)
PGI-I:
3.1(1.7) vs. 3.0(1.4) | Continuation
phase Lamotrigine vs. Placebo CGI-S, mean change from
baseline: -2.2(1.3) vs. -2.2(1.3)
CGI-I: 1.8(1.0) vs.
1.8(0.9)
MADRS, mean change from baseline:
−20.9(10.1) vs. −21.6(9.0) | Continuation
phase Lamotrigine vs. Placebo QLDS, mean change from
baseline: − 15.6(9.0) vs. −
15.2(10.7)
PGI−I: 1.9(1.2) vs. 1.9(0.9) | Monitored | Double-blind phase;
[Continuation phase] Lamotrigine
vs. Placebo AEs: 97(75%) vs. 84(71%);
[115(74%)] Headache:
32(25%) vs. 26(22%);
[17(11%)] Infection:
14(11%) vs. 25(21%);
[20(13%)] Nausea:
11(9%) vs. 18(15%);
[8(5%)] Rash: 12(9%)
vs. 7(6%);
[12(8%)] Xerostomia:
12(9%) vs. 5(4%);
[0(0%)] Back pain:
7(5%) vs. 6(5%);
[5(3%)] Insomnia:
7(5%) vs. 5(4%);
[9(6%)] Influenza:
6(5%) vs. 6(5%);
[0(0%)] Abdominal pain:
6(5%) vs. 2(2%) Diarrhea: 6(5%)
vs. 4(3%);
[5(3%)] Pain: 6(5%)
vs. 4(3%);
[6(4%)] Accidental injury:
5(4%) vs. 6(5%);
[13(8%)] Dizziness:
4(3%) vs. 8(7%);
[5(3%)] Vomiting:
2(2%) vs. 7(6%);
[0(0%)] Pharyngitis:
0(0%) vs. 0(0%);
[7(5%)] Fatigue:
0(0%) vs. 0(0%);
[5(3%)] Somnolence:
0(0%) vs. 0(0%);
[5(3%)] Serious AEs:
5(4%) vs. 6(5%);
[9(6%)] Suicidal:
2(2%) vs. 2(2%);
[0(0%)] Suicidal ideation:
0(0%) vs. 0(0%);
[3(2%)] Attempted suicide:
1(1%) vs. 1(1%);
[0(0%)] Mania: 1(1%)
vs. 0(0%);
[5(3%)] Dystonic movement:
0(0%) vs. 1(1%);
[0(0%)] Meningitis:
0(0%) vs. 1(1%);
[0(0%)] Motor dysfunction:
0(0%) vs 1(1%);
[0(0%)] | Double-blind
phase Lamotrigine vs. Placebo Total withdrawals:
52/133(39%) vs. 33/124(27%) Withdrawals due
to AEs: 16/133(31%) vs.
9/124(27%)
Continuation
phase Total withdrawals:
56/161(35%) Withdrawals due to AEs:
13/161(23%) | |
Solomon, 1997 U.S. (Poor) | Pilot long-term, double-blind, placebo-controlled
RCT Inpatient then outpatient setting | Current episode of mania or major depression;
bipolar I disorder (DSM III-R); > 1 mood episode in previous 3 y; age
18 to 65 y | Divalproex (titrated to serum concentration of 50
to 125 μg/ml) vs. Placebo for up to 12 mo. Both agents in
combination with lithium (titrated to serum concentration of 0.8 to 1.0
mmol/l) | Run-in on treatment directed at controlling the
acute episode (details not reported); patients were randomized once subjects
began to show signs of improvement from the index episode | Neuroleptics, antidepressants,
benzodiazepines | Modified version of the Longitudinal Interval
Follow-up Evaluation (LIFE), recorded at baseline and every 2 mo. This
included a 6- point Psychiatric Status Rating (PSR) scale (1 =
no symptoms, 6 = symptoms that meet full criteria for a
DSM-III-R disorder along with psychosis or extreme impairment in
functioning).
Partial remission
= improvement, but continued moderate to marked symptoms not
meeting full criteria for a mood episode (PSR of 3 or 4).
Relapse = return of symptoms that met
DSM-III-R criteria for a definite mood episode (PSR of 5 or 6) and occurred
during a period of partial remission. Recovery
= at least 8 consecutive weeks of no symptoms or minimal
symptoms (PSR of 1 or 2, respectively). Recurrence
= reappearance of the DSM-III-R disorder at full criteria (PSR
of 5 or 6) after recovery from the preceding episode (i.e., new mood
episode). | Divalproex (+Lithium) vs. Placebo
(+Lithium) Age, range, y: 31 to 65 vs. 30 to
41 Male/Female: 4/1 vs. 4/3 Ethnicity: Not reported | Number of lifetime mood episodes, range: 2 to 51
vs. 3 to 30 (mean data not reported; NSD) Past lithium treatment, n
(%): 1/5 (20.0%) vs. 6/7
(85.7%) Major depression at intake, n (%):
4/5 (80.0%) vs. 2/7 (28.6%) (NSD) Mania
episode at intake, n (%): 1/5 (20.0%) vs. 5/7
(71.4%) (NSD) | Numbers screened and eligible not reported/12
enrolled/12 randomized | 4 withdrew/None lost to follow-up/12 analyzed | Divalproex vs. Placebo
Partial
remission, n: 5/5 (100%) vs. 6/7 (85.7%) (1
divalproex patient recovered prior to randomization; 1 placebo patient
recovered abruptly in wk 4 with no intervening period of partial
remission) Time to partial remission, range, wk: 0 to 1 vs. 1 to
11
Relapse or recurrence, n (%): 0/5
(0.0%) vs. 5/7 (71.4%) (p = 0.014) | | | | Monitored | Most common adverse events on divalproex
(+ lithium): distress, tremor, cognitive impairment,
alopecia Adverse events on placebo (+ lithium): not
reported | gastrointestinal Total withdrawals: 2/5
(40.0%) vs. 2/7 (28.6%) Withdrawals due to
adverse events: 2/5 (40.0%) vs. 0/7 (0.0%) | Results are inconclusive (pilot study). Small
sample size, confounding co-medications, nonblinded research
psychiatrist. |
Calabrese, 1999 Australia, France, U.K.,
U.S. (Fair) | Multicenter, double-blind, double-dummy,
placebo-controlled, parallel-group RCT Outpatient setting | Bipolar I disorder (DSM-IV); at least 2 previous
mood episodes in past 10 years with at least 1 episode a manic or mixed
episode; current major depressive episode of >/= 2 wk but
</= 12 months in duration; minimum score of 18 on 17-item
Hamilton Rating Scale for Depression (HAM-D) | Lamotrigine titrated to 50 mg/d (at target dose
from wk 3 to 7) vs. Lamotrigine titrated to 200 mg/d (at target dose from wk
5 to 7) vs. Placebo for 7 wk | Washout of previous psychoactive drugs within a
time equivalent to 5 elimination half-lives prior to randomization | Chloral hydrate, lorazepam,
temazepam. oxazepam during first 3 wk of treatment | HAM-D, Montgomery-Asberg Depression Rating Scale
(MADRS); Mania Rating Scale (MRS), Clinical Global Impressions scale for
Severity (CGI-S) at baseline and weekly for 7 wk, and Clinical Global
Impressions scale for Improvement (CGI-I) from day 4
onward.
Response was defined as 50% or more
reduction on the 17-item HAM-D or MADRS scales or a rating of very much
improved or much improved on the CGI-I scale. | Lamotrigine 50 mg/d (N = 66) vs.
Lamotrigine 200 mg/d (N = 63), vs. Placebo (N =
66) Age, mean, y: 41 vs. 42, vs. 42 Male / Female:
33% / 67% vs. 44% / 56% vs.
41% / 59% Ethnicity not reported | Age of onset of affective symptoms, mean, y: 22
vs. 21 vs. 21 No. of mood episodes in last 12 mo per patient, mean
(SD): 2.2 (0.8) vs. 2.2 (0.9) vs. 2.2 (0.8) Duration of current
episode --2 to 8 wk: 39% vs. 37% vs.
29% --> 8 to 24 wk: 44% vs.
41% vs. 42% --> 24 wk: 17%
vs. 22% vs. 29% Moderate intensity of
depression: 58% vs. 54% vs.
61% CGI-S score (% of
patients) --Mildly ill: 3% vs. 10% vs.
2% --Moderately ill: 64% vs. 51%
vs. 65% --Markedly ill: 23% vs.
30% vs. 28% --Severely ill: 11%
vs. 10% vs. 11% Melancholic features:
39% vs. 40% vs. 50% Prior
hospitalization for mood episode: 44% vs. 51% vs.
62% Prior suicide attempts: 32% vs.
32% vs. 36% Lithium use in last 5 mo:
23% vs. 19% vs. 23% | Numbers screened, eligible, and enrolled not
reported / 195 randomized | 60 withdrew / None reported / 192 analyzed for
efficacy, 194 analyzed for safety | Lamotrigine 50 mg/d (N = 64) vs.
Lamotrigine 200 mg/d (N = 63) vs. Placebo (N = 65)
(Last observation carried forward [LOCF]
analysis) Change in scores from baseline,
mean 17-item HAM-D (Primary efficacy variable):
−9.3 vs. −10.5 vs. −7.8 (p =
0.084) (Analysis for observed change showed a significant treatment
difference in change from baseline: −12.6 (N = 43)
vs. −13.2 (N = 45) vs. −9.3 (N
= 47) (p < 0.05 for both lamotrigine groups vs.
placebo) Significant improvement was first noted for lamotrigine 200
mg/d only vs. placebo at week 5 (p < 0.05). | Change in scores from baseline, mean MADRS:
−11.2 vs. −13.3 vs. −7.8 (p <
0.05 for lamotrigine 200 vs. placebo) CGI-S: −1.0 vs.
−1.2 vs. −0.7 (p < 0.05 for lamotrigine 200
vs. placebo) CGI-I: 3.0 vs. 2.6 vs. 3.3 (p < 0.05 for
lamotrigine 200 vs. placebo) MRS: 0.9 vs. 0.3 vs. −0.5
(NSD) | Combined week 3 analysis (lamotrigine
</= 50 mg/d for both active groups) (N =
127): significant improvements (p < 0.05) were seen by week 3 in
HAM-D Item 1 and MADRS for LOCF analyses. Subgroup analysis: No
significant effect of recent lithium use on treatment group differences for
any efficacy measure. | Responder rate 17-item
HAM-D: 45% vs. 51% vs. 37%
(NSD) MADRS: 48% vs. 54% vs. 29%
(p < 0.05 for each lamotrigine group vs. placebo) CGI-I:
41% vs. 51% vs. 26% (p < 0.05 for
lamotrigine 200 vs. placebo) | Elicited by investigator | Lamotrigine 50 mg/d (N = 66) vs.
Lamotrigine 200 mg/d (N = 66) vs. Placebo (N =
65) Patients reporting any adverse event: 79% vs.
79% vs. 92% Of the most common
(>/= 5%) adverse events, only headache showed
a significant treatment difference (n, %): 23 (35%)
vs. 20 (32%) vs. 11 (17%) (p < 0.05 for each
lamotrigine group vs. placebo) Other common adverse
events: --Nausea: 11 (17%) vs. 10 (16%) vs.
10 (15%) --Pain: 5 (8%) vs. 7
(11%) vs. 5 (8%) --Rash: 9 (14%)
vs. 7 (11%) vs. 7 (11%) --Dizziness: 6
(9%) vs. 6 (10%) vs. 2 (3%) Manic
/ hypomanic / mixed episodes (as reported by investigator) (n,
%): 2 (3%) vs. 5 (8%) vs. 3
(5%) (NSD)
Patients reporting any serious adverse event: 4 vs. 2 vs. 3 Illness-related Serious Adverse
Events --Probable suicide: 0 vs. 0 vs. 1 --Attempted suicide:
1 vs. 0 vs. 1 --Suicidal ideation: 1 vs. 1 vs. 0 --Worsening
depression: 1 vs. 0 vs. 0 --Psychotic episode: 1 vs. 0 vs.
0 (All illness-related serious adverse events in the lamotrigine
50-mg/d group except for the attempted suicide [3 out of 4
events] were considered to be possibly drug related.) | Lamotrigine 50 mg/d vs. Lamotrigine 200 mg/d vs.
Placebo Total withdrawals: 23 (35%) vs. 18
(29%) vs. 19 (29%) Withdrawals due to adverse
events: 12 (18%) vs. 10 (16%) vs. 10
(15%) Adverse events accounting for more than one
withdrawal --Rash: 3 vs. 4 vs. 2 --Worsening of psychiatric
depression: 3 vs. 0 vs. 1 --Pruritus: 0 vs. 1 vs. 1 --Suicidal
ideation: 1 vs. 1 vs. 0 --Suicide attempt: 1 vs. 0 vs.
1 --Mania: 0 vs. 2 vs. 0 | Modified ITT analyses were used for efficacy and
safety. Dosage escalation was faster than the recommended regimen and may
have increased the risk of rash. The fixed-dose titration schedule resulted
in unequal treatment durations for the 50-mg group (5 wk) and the 200-mg
group (3 wk). The 17-item HAM-D scale (weighted toward somatic
symptomatology) may have been less sensitive and reliable for detecting
effects on bipolar depression or treatment differences than the MADRS. |
Calabrese, 2000 U.S.,
Canada (Fair) | Multicenter, double-blind, placebo-controlled,
parallel- group RCT Outpatient setting implied | Age 18 y or older; bipolar disorder I or II with
rapid cycling (DSM-IV); euthyroid or, if taking thyroid replacement therapy,
on stable dose for 3 mo | Open-label preliminary phase: Lamotrigine
started at 25 mg/d and slowly titrated to target dose of 200 mg/d (max. 300
mg/d) for 4 to 8 wk
Double-blind phase: Lamotrigine
100 to 500 mg/d vs. Placebo for 26 wk
Lamotrigine doses were
adjusted for concomitant valproate or carbamazepine therapy. | 4- to 8-wk run-in on lamotrigine; patients were
randomized if they were taking a minimum dose of 100 mg/d of lamotrigine and
had a score of </= 14 on the 17-item Hamilton Rating
Scale for Depression (HAM-D) and </= 12 on the Mania
Rating Scale (MRS) from the Schedule for Affective Disorders and
Schizophrenia (SADS)-Change version over a 2-wk period; they were eligible
to enter the randomized phase if they successfully completed a taper of all
other psychotropic medications while maintaining the minimum criteria for
wellness, had no change in lamotrigine dosage during the final week of the
preliminary phase, and had no mood episodes requiring additional drug or
electroconvulsive therapy after the first 4 wk of the preliminary phase | Open-label phase: Lithium (60,
19%), divalproex (63, 19%), carbamazepine (14,
4%), antidepressants (96, 30%), antipsychotics (24,
7%), and benzodiazepines (88,
27%) Double-blind phase: Lorazepam. Other psychotropics
(e.g., lithium, divalproex, antipsychotics, electroconvulsive therapy) could
be added only if an increase in lamotrigine dose was not effective or
appropriate (i.e., patients reached primary study end point). | Open-label phase: 17-item HAM-D, MRS, Clinical
Global Impressions-Severity scale (CGI-S), Global Assessment Scale (GAS),
and retrospective life chart at screening (within −14 d), day 1,
then weekly till randomization.
Double-blind phase: HAM-D,
MRS, CGI-S, GAS, and prospective life chart on day 1, then wk 2, 3, 4, 6, 8,
10, 12, 16, 20, 24, and 26.
Relapse was
operationally defined as the need for additional pharmacotherapy for a mood
episode or one that was thought to be emerging. | Open-label Lamotrigine (N =
324); Double-blind Placebo (N = 88) vs. Lamotrigine (N
= 92) Age, mean, y: 38.6; 37.4 vs. 38.5 Female, n
(%): 190 (59%); 52 (59%) vs. 51
(55%) Ethnicity: Not reported | Age at onset of first episode of depression /
mania, mean, y: 17.5 / 20.2; 17.0 / 19.1 vs. 17.3 / 20.7 Bipolar I, n
(%): 225 (69%); 60 (68%) vs. 68
(74%) Bipolar II, n (%): 98
(30%); 28 (32%) vs. 24 (26%) No.
of mood episodes in last 12 mo, mean: 6.3; 5.9 vs. 6.3 Prior
hospitalizations for mood episode, mean: 1.8; 1.3 vs. 1.5 Prior
suicide attempt, n (%): 117 (36%); 34
(39%) vs. 25 (27%) Lifetime prevalence of
psychosis, n (%): 88 (27%); 21 (24%) vs.
25 (27%) Type of mood episode at screening,
% --Depression: 57%; 56% vs.
55% --Mania/Hypomania: 20%; 19%
vs. 20% --No episode: 18%; 17%
vs. 21% --Mixed: 5%; 9% vs.
4% | Numbers screened and eligible not reported / 324
enrolled / 182 randomized | Open-label phase: 142 withdrew / 19 lost to
follow-up / 324 analyzed for safety
Double-blind phase: 28
withdrew / 10 lost to follow-up / 177 analyzed for efficacy, 180 for
safety | Lamotrigine vs. Placebo Time to relapse
(Primary Efficacy Measure), median survival time, wk: 18 vs. 12 (p
= 0.177) --In bipolar I subgroup (N = 125):
18 vs. 14 (estimated; p = 0.738) --In bipolar II subgroup
(N = 52): 17 vs. 7 (p = 0.073) Required
additional pharmacotherapy for emerging mood episode, n (%): 45
(50%) vs. 49 (56%) | Time to premature discontinuation for any reason,
median survival time, wk: 14 vs. 8 (p = 0.036) --In
bipolar I subgroup: 10 vs. 12 (estimated; p = 0.426) --In
bipolar II subgroup: 16 vs. 5 (estimated; p =
0.015)
Stable without relapse for 6 mo, n (%):
37/90 (41%) vs. 23/87 (26%) (p =
0.03) --In bipolar I subgroup: 39% vs. 31%
(NSD) --In bipolar II subgroup: 46% vs. 18%
(p = 0.04) | CGI-S, change from baseline: NSD (data not
reported) --In bipolar I subgroup: NSD --In bipolar II
subgroup: NSD
GAS, change from baseline: NSD (data not
reported) --In bipolar I subgroup: NSD --In bipolar II
subgroup: p </= 0.03 at wk 3, 6, and
12
17-item HAM-D, change from baseline: NSD (data not
reported) MRS, change from baseline: NSD (data not reported) | | Monitored | Double-blind phase--Lamotrigine (N =
92) vs. Placebo (N = 88) Serious adverse events, n: 1 vs.
2 Adverse events considered reasonably related to study treatment: 24
(27%) vs. 28 (30%) (NSD); most common: nausea (4,
4% vs. 4, 5%) and headache (6, 7% vs. 8,
9%) Most Common (>/= 10%)
Treatment-emergent Adverse Events: headache (21, 23% vs. 15,
17%), nausea (13, 14% vs. 10, 11%),
infection (11, 12% vs. 10, 11%), pain (9,
10% vs. 7, 8%), and accidental injury (10,
11% vs. 4, 5%). Rash occurred in 3
(3%) vs. 2 (2%) patients. Treatment-related
rash: 0 (0%) | Double-blind phase Total withdrawals: 11/93
(12%) vs. 17 (19%) Withdrawals due to adverse
events: 1 (1%) vs. 2 (2%) | The analyses for double-blind treatment were based
on a selective cohort of patients who were more likely to be lamotrigine
responders and less prone to develop rash. The primary efficacy measure,
time to relapse, depended on the investigator's discretion of
whether additional psychotropic medication was necessary to treat an
emerging mood episode. |
Mishory, 2003 Israel (Poor) | Double-blind, placebo-controlled, crossover
RCT Outpatient setting | Bipolar disorder I or schizoaffective disorder
(DSM-IV); no unstable physical illness; out of hospital for at least 1 mo;
inadequate prophylaxis in the past on lithium, carbamazepine, or valproate;
at least 1 episode per year for previous 2 years despite compliance with
their mood stabilizer | Phenytoin (starting at 100 mg and titrated by 100
mg/wk; mean dose and serum concentration at 6 mo: 380
+/− 80 mg and 10.7 +/− 4.2
mcg/ml) vs. Placebo for 6 mos then crossover | 1-mo phased washout during crossover | Ongoing prophylactic treatment remained unchanged
(lithium, carbamazepine, valproate, or neuroleptic) | Brief Psychiatric Rating Scale (BPRS), Young Mania
Scale (YMS), Hamilton Depression Scale (HMS), and Global Clinical Impression
at baseline and monthly thereafter
Primary outcome measure
was time to 'event,' an affective relapse. Criteria for an
'event' were need for hospitalization or emergent symptoms
of sufficient severity to require addition of a neuroleptic or
antidepressant, according to the masked clinical psychiatrist. | Age. mean (SD), y: 45.2 (9.6) Male /
Female: 9 / 14 Ethnicity not reported | Age of onset of illness, mean (SD), y: 26.5
(9.0) Number of affective episodes, mean (SD): 13.8 (8.5) Time
in remission before entering trial, mo: 4.0 (range: 1 to 13) Last
affective episode --Mania: 11 --Depression: 7 --Mixed:
5 | Number screened, eligible, enrolled not reported /
23 randomized | 4 withdrew (and were replaced with new enrolled
patients) / None lost to follow-up / 23 analyzed (30 6- mo observation
periods) | Phenytoin vs. Placebo Time to clinical
relapse (event), median (estimated from figure), mo: > 6 vs. 5 (p
= 0.02) Relapsed during first 6 mo: 3/10
(30.0%) vs. 8/13 (61.5%) (p =
0.053) Data for rating scales were not reported. | | | | Not reported | Phenytoin (n = 14) vs. Placebo (n
= 16) Common adverse events during 30 observation
periods Slight weakness and sleepiness: 1 (7.1%) vs. 1
(6.2%) Temporary dizziness, resolved without change in
treatment: 3 (21.4%) vs. 0
(0.0%) Psoriasis-like symptoms: 1 (7.1%) vs.
0 (0.0%) | Phenytoin vs. Placebo Total withdrawals:
9/23 (39.1%) vs. 7/23 (30.4%) (if 4 dropouts during
the first 3 wk of phenytoin treatment are counted, total for phenytoin would
be 13/27, 48.1%) Withdrawals due to adverse event: 1/23
(4.3%) vs. 0/23 (0.0%) (psoriasis-like symptoms due
to concomitant lithium treatment) | Small sample size; dropouts excluded from
analyses; short study duration; incomplete reporting of data. Results
reflected a selective population of compliant patients because any
post-randomization dropout was excluded from analyses and replaced with a
new patient who was assigned the dropout's randomization
number. |
Pande, 2000 U.S. (Fair) | Multicenter, double-blind, parallel-group
RCT Outpatient setting | Age 16 y or older; lifetime diagnosis of bipolar I
disorder (DSM-IV) with manic/hypomanic or mixed symptoms; Young Mania Rating
Scale (YMRS) >/= 12 despite ongoing treatment with
lithium, valproate, or both in combination; lithium serum concentration
>/= 0.5 mEq/l or valproate concentration
>/= 50 mcg/ml | Gabapentin 600 to 3600
mg/d Placebo 10 wk (Added on to lithium, valproate, or
combination) | 2-wk, single-blind, placebo run-in during which
lithium and/or valproate doses were adjusted based on clinical response and
to achieve minimum threshold concentrations; patients were randomized to
double-blind treatment if they met entry criteria at the end of the placebo
run-in | Lithium and valproate at steady doses unless
dosage changes were necessary for patient safety | YMRS, Hamilton Depression Rating Scale (HAM-D),
Hamilton Anxiety Rating Scale (HAM-A), Clnical Global Impression of Severity
(CGI-S) and Change (CGIC), recorded weekly for 4 wk after randomization,
then biweekly for 6 wk. Self-assessed internal state scale (ISS), Life Chart
for Recurrent Affective Illness (Life Chart), and SF-36 Quality of Life
Questionnaire
Responders were defined as "much improved" or
"very much improved" on CGIC | Gabapentin (N = 58) vs. Placebo (N
= 59) Age, mean (SD), y: 40.7 (.4) vs. 38.2
(10.5) Male / Female, %: 50 / 50 vs. 54 /
46 Ethnicity not reported | Ongoing treatment for bipolar
disorder --Lithium only, n: 22 vs. 17 --Valproate only, n: 26
vs. 31 --Both, n: 10 vs. 11 | Numbers screened and eligible not reported / 117
enrolled / 117 randomized | 48 withdrawn / None lost to follow-up / 114
analyzed | Gabapentin vs. Placebo Adjusted means
included treatment and center in ANCOVA model and YMRS baseline score as
covariate YMRS, adjusted mean: −6.5 vs. −9.9
(difference −3.34; 95% CI: −6.35 to
−0.32; p = 0.03) HAM-D, adjusted mean: 0.01
vs. −1.3 (difference −1.32; 95% CI:
−4.40 to 1.77; p = 0.40) | Change in score from baseline to last observation
carried forward HAM-A, total score: 0.36 vs. −1.05 (p
= 0.24) CGI-S: −0.63 vs. −0.98 (p
= 0.10)
ISS, % of
patients --Manic (>/= 70): 9 vs.
8 --Depressed (</= 30): 17 vs. 17 --Normal
(31 to 69): 74 vs. 75 | CGIC "much improved" or "very much improved"
(responders), % : 37 vs. 47 (p = 0.30) | | Monitoring | Gabapentin vs. Placebo
Serious
adverse events: 6 vs. 5 (3 of the 6 serious adverse events in the gabapentin
group started during the placebo lead-in)
Most frequent
adverse events, % --Somnolence: 24.1 vs.
11.9 --Dizziness: 19.0 vs. 5.1 --Diarrhea: 15.5 vs.
11.9 --Headache: 10.3 vs. 11.9 --Amnesia: 10.3 vs. 3.4 | Gabapentin vs. Placebo Total Withdrawals:
27/58 (46.6%) vs. 21/59 (35.6%) Withdrawals
due to adverse events: 7/58 (12.1%) vs. 5/59
(8.5%) | Primary efficacy variables were the YMRS and
HAM-D. Placebo was superior to gabapentin in terms of changes in YMRS
scores. A post hoc analysis determined that more lithium dosage adjustments
were made during the placebo lead-in in the placebo group (n =
12) than in the gabapentin group (n = 4; p < 0.01). When
the data from these 16 patients were excluded from analysis, the treatment
difference in YMRS change score was no longer significant. |
Weisler, 2004, Shire Dossier,
2005 U.S. SPD417 Study (Fair) | Multicenter (24 sites) double-blind,
placebo-controlled, parallel-group RCT Inpatient then outpatient
setting | Age at least 18 y; bipolar I disorder with current
manic or mixed episodes (DSM-IV); history of at least 1 previous manic or
mixed episodes; minimum screen and baseline total score of 20 on Young Mania
Rating Scale (YMRS); enrollment of treatment-resistant patients was
discouraged | Carbamazepine extended-release capsules (CBZ ERC)
started at 400 mg/d then titrated based on investigator discretion to 200 to
1600 mg/d vs. Placebo for 4 wk --Mean final daily dose of CBZ ERC:
756 mg --Median final dosage range (N=192, ITT): 800 to
1000 mg --Mean plasma drug concentration: 8.9 mcg/ml | Single-blind placebo lead-in for first 7 days | Lorazepam, acetaminophen, and ibuprofen; other
less commonly used allowed co-medications were not reported | YMRS, Clinical Global Impression of Severity
(CGI-S) and Improvement (CGI-I) scales; Hamilton Rating Scale for Depression
(HAM-D), adverse events, and adherence, every week; physical examination,
hematology, blood chemistry, and urinalysis at screening, baseline, and
termination visit
Responder rate defined as percentage of
patients with at least 50% decrease in YMRS scores from baseline
to last observation | CBZ ERC (N = 101) vs. Placebo (N
= 103) Age, mean, y: 38.0 vs. 38.1 (NSD) Female,
n: 41 (40.6%) vs. 56 (54.4%) (p =
0.0489) White, n: 73 (72.3%) vs. 75 (72.8%)
(p = 0.2924) | Mixed episode, n: 60 (59.4%) vs. 48
(46.6%) (p = 0.0670) | Numbers screened, eligible, enrolled not reported
/ 204 randomized | Of 204 randomized: 108 (52.9%)
withdrew / 6 lost to follow-up / 192 analyzed (ITT) | CBZ ERC (N = 94) vs. Placebo (N
= 98)
YMRS total score, mean --Baseline:
27 vs. 28 --Day 21, primary end point (Calculated change from
baseline): 18 (−8.70) vs. 23 (−5.17) (calculated
difference, −4; p < 0.033) --First statistically significant difference seen at day 14
Responder
rate --Day 21: 69% vs. 30% (p <
0.003) Calculated NNT: 3 (2 to 4) --End point:
41.5% vs. 22.4% (p <
0.0074) Calculated NNT: 5 (3 to 16) --First statistically significant difference seen at day 14 | Subgroup analyses YMRS total
score --By gender, 3 age groups, white vs. nonwhite, manic vs. mixed
episode: similar moderate treatment effects in favor of CBZ ERC in all
subgroups Change in YMRS total score from baseline to end
point --Manic episode patients: −6.44 vs.
−1.8 (p = 0.0092) --Mixed episode patients:
−10.31 vs. −9.8 (NSD) | CGI-S score, change (improvement) from baseline to
end point / day 21: 4.07 vs. 3.66 (p = 0.0254) CGI-I
score, mean % change at day 21: 66.7% vs.
35.3% (p = 0.0035) CGI-I score, mean
% change at end point: 43.6% vs. 24.0%
(p = 0.0067)
HAM-D score, mean change from
baseline to day 21: −5.35 vs. −1.58 (p =
0.09) Post hoc subgroup analysis of change in HAM-D score from
baseline in mixed-episode patients remaining on CBZ ERC treatment at day 21:
−7.62 vs. −2.44 (p = 0.01) | Took allowed co-medication: 89.1% vs.
90.3% --Lorazepam: 71.3% vs.
67.0% (NSD) --Lorazepam dose (n = 83), mg:
2.2 vs. 2.2
Daily adherence rate, mean: 92.4% vs.
93.4% | Monitoring | CBZ ERC (N = 101) vs. Placebo (N
= 103) Serious AEs, n: 4 (4.0%) vs. 4
(3.9%) --Worsening/Exacerbation of bipolar symptoms, n: 4
vs. 3 --Suicidality with rehospitalization, n: 0 vs.
1 --Deaths: None Total AEs, n: 89 (88.1%) vs. 75
(72.8%) (p = 0.0078) Possibly related /
related AEs, n: 78 (77.2%) vs. 59 (57.3%) (p
= 0.0029)
Notable Treatment-emergent AEs with a
significant treatment difference, n --Dizziness: 49
(48.5%) vs. 13 (12.6%) --Nausea: 38
(37.6%) vs. 11 (10.7%) --Somnolence: 33
(32.7%) vs. 16 (15.5%) --Vomiting: 22
(21.8%) vs. 4 (3.9%) --Dyspepsia: 19
(18.8%) vs. 5 (5.8%) --Dry mouth: 12
(11.9%) vs. 3 (2.9%) --Pruritus: 9
(8.9%) vs. 2 (1.9%) --Speech disorder: 7
(6.9%) vs. 0 (0.0%)
Other selected
AE, n --Rash: 9 (8.9%) vs. 6 (5.8%)
(NSD) | CBZ ERC (N = 101) vs. Placebo (N
= 103) Total withdrawals: 51 (50.5%) vs. 57
(55.3%) (NSD) Withdrawals due to serious AEs: 3
(treatment group(s) not reported) Withdrawals due to AEs: 13
(12.9%) vs. 6 (5.8%) (p =
0.0959) --Nausea, dizziness, mania, pruritus: each 2
(2.0%) vs. 0 (0.0%) --Rash: 2
(2.0%; 1 severe) vs. 2 (1.9%) --Diarrhea: 0
(0.0%) vs. 2 (1.9%)
Laboratory
results showing significant treatment differences --Alkaline
phosphatase, mean absolute (relative %) change, U/l: 8.035
(12%) vs. 1.686 (2%) (p =
0.0108) --Cholesterol, mean change, mg/dl: 21.4 vs. 1.1 (p <
0.0001) --White blood cell count, mean change (final value),
103/μl: 1 151 vs 0 053 (p | Subgroup analysis of change in YMRS scores showed
statistically significant treatment difference only in manic patients
because of a greater placebo response in mixed-episode patients. Authors
note that an antidepressant effect would not be expected to occur in a 3-wk
trial. Trial was not powered to detect rare AEs, such as agranulocytosis
(1.4 per 1 million patients treated per year) and aplastic anemia (5.1 per 1
million patients treated per year). |
Weisler, 2005 U.S., India SPD417
Study (Fair) | Multicenter, double-blind, placebo-controlled,
parallel- group RCT Inpatient then outpatient (after day 7 of
double-blind treatment, patient could be discharged at
physician’s discretion) | Age > / = 18 y; DSM-IV criteria
for bipolar I disorder with current manic or mixed episodes; history of at
least one previous manic or mixed episode; minimum prestudy and baseline
Young Mania Rating Scale (YMRS) total score of 20 | Carbamazepine extended- release capsules (CBZ ERC)
started at 400 mg/d then titrated based on investigator discretion to 200 to
1600 mg/d vs. Placebo for 21 d (double- blind treatment phase) then
30-d follow-up (for safety) --Most patients titrated to final daily
dose of CBZ ERC 400 to 1000 mg | 5-day single-blind placebo run-in to ensure
washout of previous bipolar treatment and exclusionary medications | Lorazepam--through, and not after, the second week
of double- blind treatment | YMRS, Clinical Global Impression of Severity
(CGI-S) and Improvement (CGI-I) scales, Hamilton Rating Scale for Depression
(HAM- D), time to outpatient status, physical examination,
electrocardiogram, laboratory assessments, adverse event
reporting
Responder rate was the percentage of patients with
> / = 50% decrease (improvement) in YMRS
scores from baseline to last observation | Carbamazepine ERC (N = 122)
vs. Placebo (N = 117) Age, mean, y:
37 Male,%: 70% From U.S.:
62% From India: 38% Caucasian:
46% Manic episode: 79.1% | Mixed episodes: 21% Received
prior bipolar treatment: 90% | Numbers screened, eligible, enrolled not reported
/ 239 randomized | 95 (39.7%) withdrew / 4 lost to
follow-up / 235 analyzed | CBZ ERC (N = 120) vs. Placebo (N
= 115) Mean change from baseline to day 21 --YMRS
total score: −15.1 vs. −7.1 (p <
0.0001) --CGI-S score (improvement): 1.5 vs. 0.6 (p <
0.0001) --HAM-D total score: −2.7 vs. −1.0 (p
= 0.008) --HAM-D depressed mood item number 1 score: NSD
at any time point (data not reported)
Responder rate: 73/120
(61%) vs. 33/115 (29%) (p <
0.0001) Calculated NNT: 3 (2 to 5) | Outpatient status: 48.3% vs.
38.4% (p < 0.05)
Time to discharge: 14.1 d
in both groups
Onset (time to first statistically significant
effect): 7 d Withdrawals due to lack of efficacy: 6.6%
vs. 23.1% (p = 0.0004) | Subgroup analyses by age, gender, country, manic
or mixed episode --YMRS total scores: similar decreases (data not
reported) --HAM-D: significant treatment difference in manic subgroup
(p < 0.05); NSD in mixed episode subgroup (p =
0.0607) | Concomitant medications: 91.8% vs.
86.3% (mostly lorazepam, ibuprofen,
acetaminophen)
Concomitant lorazepam: 73.8% vs.
78.6% | Monitoring | CBZ ERC (N = 122) vs. Placebo (N
= 117) Serious AEs: 3.3% vs. 5.1%
(NSD) --One SAE was considered to be possibly related to study
treatment: fever, erythematous macular rash over trunk and lower extremities
and low white blood cell count --No deaths
Any
treatment-emergent AE: 91.8% vs. 56.4% (p <
0.0001) AEs occurring at a significantly higher rate on CBZ ERC than
Placebo: dizziness, somnolence, nausea, ataxia, vomiting, and blurred
vision --Dizziness: 39.3% vs. 12.0% (p
< 0.0001) --Somnolence: 30.3% vs.
10.3% (p = 0.0001)
Other selected
AEs: --Rash: 4.9% vs. 2.6%
(NSD) --Pruritus: 8.2% vs. 2.6%
(NSD)
Percent change from baseline to end point --WBC
count: −11.7% vs. 0.3%
(p=0.0001) --Total cholesterol: 13.2% vs.
2.0% (p<0.0001) --Low-density lipoprotein (LDL):
28.1% vs. 11.5%
(p<0.0001) --High-density lipoprotein (HDL): 9.7%
vs. 3.2% (p<0.01)
Clinically significant
increase in LDL, n: 1 vs. 0 Clinically significant increase in
triglycerides, n: 1 vs. 0 | Total withdrawals: 34.4% vs.
45.3% (NSD) Withdrawals due to AEs: 9.0% vs.
5.1% (NSD) | All patients were hospitalized during the run-in
period and for at least the first 7 days of double- blind treatment, after
which patients could be discharged if stable. |
Salloum,
2005 U.S.
(Fair) | Two-center double-blind, placebo-controlled,
parallel- group RCT Outpatient setting implied | Age 18 to 65 y; after clearing of acute withdrawal
symptoms (using Revised Clinical Institute Withdrawal Assessment for
Alcohol Scale), met 4 of 7 DSM-IV alcohol dependence criteria; actively
drank alcohol in past month; concurrent acute episode of bipolar I disorder
(manic, mixed, or depressed) | Divalproex started at 750 mg/d then titrated to
serum concentration of 50 to 100 mcg/ml (mean, 51.5 mcg/ml) vs. Placebo for
24 wk (as add-on to lithium) | None | Lithium (to trough concentration of 0.7 to 1.2
mEq/l); perphenazine; benztropine; sertraline; trazodone; dual diagnosis
recovery counseling; participation in self- help groups (e.g., Alcoholics
Anonymous; dual Recovery Anonymous; manic- depressive support group) | Timeline Follow-back for Recent Drinking; Modified
Quantitative Alcohol Inventory / Craving Scales; Weekly Self-Help Activity
Questionnaire; Somatic Symptoms Checklist; Medication Adherence Form; breath
alcohol concentration, urine drug screen; number of drinks consumed;
proportion of heavy drinking days ( > / = 4 drinks/d for
women; > / = 5 drinks/d for men); number of drinks per
heavy drinking day; time to relapse to sustained heavy drinking (3
consecutive heavy drinking days); Hamilton Rating Scale for Depression
(HRSD-25); Bech-Rafaelsen Mania Scale (BRMS); Global Assessment Scale (GAS);
remission of mania (score of </=7 on BRMS); remission of
depression (score of </=7 on HRSD-25) every 2 wk for 24
wk | Divalproex vs. Placebo Age, mean, y:
37 vs. 38 Male, n: 21 (72%) vs. 23
(77%) African American, n: 8 (28%) vs. 7
(23%) | Mixed bipolar, n: 30
(58%) Manic: 11 (21%) Depressed: 11
(21%) Attempted suicide during index episode: 6
(17%) (of inpatient recruits) Other substance use
disorders, n: 26 (50%) Social class V, n: 13
(45%) vs. 11 (37%) Drinking to intoxication
in past 30 d, mean, d: 12.3 vs. 16.3 No. of drinks per week, mean: 88
vs. 104 HRSD-25 score, mean: 20.3 vs. 21.2 BRMS score, mean:
15.2 vs. 15.3 Global Assessment of Functioning score, mean: 38.1 vs.
38.4 Duration of bipolar disorder, mean, y: 13.0 vs. 15.6 | Numbers screened and eligible not reported / 72
enrolled / 59 randomized | 32 withdrew / 7 lost to follow-up (number lost to
follow-up for mood outcomes not calculable) / 52 analyzed (for alcohol use
outcome; not reported for mood outcome) | Alcohol Use
Outcome Divalproex (N = 27) vs. Placebo (N
= 25) Divalproex was superior to placebo in improving
drinking behavior (data not shown here)
Mood
Outcome Divalproex (N = 27) vs. Placebo(N
= 25) Overall mean scores (Mixed model estimate;
p-value) --BRMS (Mania) --baseline: 15.2 vs.
15.3 --final: 5.56 vs. 6.10 (−0.03;
NSD) --calculated change from
baseline: −9.64 vs. −9.20 --HRSD-25
(Depression) --baseline: 20.3 vs.
21.2 --final: 16.3 vs. 14.4 (0.12;
NSD) --calculated change from
baseline: −4.0 vs. −6.8 | Time to remission from mania (BRMS score <
/ = 7): 2 to 3 wk; earlier with divalproex but time not reported
by treatment group (p = 0.07 for difference between treatment
groups) Time to remission from depression (HRSD-25 score < /
= 7): 8 to 9 wk; not reported by treatment
group Remission from mania, n: 21 (78%) vs. 20
(80%) (calculated p = 0.86) Remission from
depression, n: 17 (63%) vs. 12
(48%) (calculated p =
0.42)
Global Assessment of Functioning
score --Baseline / Final score, mean: 38.1 / 57 vs. 38.4 /
57 --Calculated change (improvement) from baseline: 18.9 vs.
18.6 | Mixed model estimate for association between the
following: Valproate serum concentration and improvements
in --HRSD−25 scores: −0.11 (p =
0.06) --Functioning: 0.15 (p = 0.06) Manic and
depressive symptoms and alcohol use outcomes and functioning (p
= 0.006 to p < 0.001) Functioning and alcohol use
outcomes (p < 0.001) | Medication Adherence and Adjunctive
Treatment Divalproex vs. Placebo --Self-reported
medication adherence rate: 87% vs. 86%
(NSD) --Lithium serum / red blood cell concentration, mean, mEq/l:
0.68 / 0.27 vs. 0.66 / 0.32 (NSD) --Valproate serum concentration,
mcg/ml: 51.5 vs. Not reported / applicable --Participated in any
psychosocial treatment, n: 21 (78%) vs. 19
(76%) --Received adjunctive antidepressants, n: 11 / 23
(48%) vs. 10 / 21 (48%) --Received adjunctive
antipsychotics: 8 (35%) vs. 6
(29%) --Received trazodone as a hypnotic, n: 2
(9%) vs. 9 (43%) (p = 0.03) | Monitoring | Serious AEs: 0
Divalproex (N
= 27) vs. Placebo (N = 25) Treatment-emergent
AEs: NSD between treatment groups for individual AEs (not listed
here) Selected treatment-emergent AEs (NSD for any
AE) --Nausea or vomiting: 9 (39.1%) vs. 2
(9.5%) (p = 0.07) --Tremor: 11
(47.8%) vs. 14 (66.7%) --Fatigue: 7
(30.4%) vs. 10 (47.6%) --Weight gain: 3
(14.3%) vs. 5 (23.8%)
ALT and AST
levels did not differentiate between groups in mixed- model
analysis Gamma-GTP, IU/l: 66 vs. 81 (estimate, −62.08; p
= 0.045) Gamma-GTP correlated with weekly alcohol use
(estimate, 0.49; p = 0.02) | Divalproex vs. Placebo Total withdrawals:
15 (56%) vs. 17 (68%) --Required psychiatric
hospitalization: 3 / 29 (10.3%) vs. 5 / 30
(16.7%) (calculated p =
0.924) Withdrawals due to AEs: 1 (3.7%) vs. 1
(4.0%) | Authors state this is the first double-blind
placebo-controlled trial of valproate in alcoholic patients with bipolar I
disorder. Adjunctive medications and psychotherapy may have obscured
treatment differences in mood symptoms and dropout rates. Inclusion of
patients with a mixture of bipolar I mood states and a small sample size may
have reduced the study’s power to detect treatment differences
in mood symptoms. |
Davis, 2005 U.S. (Fair) | Single-center double-blind, placebo-controlled,
parallel- group RCT Outpatient setting | DSM-IV diagnosis of bipolar I disorder, currently
in depressed phase; score > / = 16 on
17-item Hamilton Rating Scale for Depression (HRSD); stable general
medicine condition; no significant abnormal laboratory values | Divalproex 500 to 2500 mg/d titrated to serum
concentration of 50 to 100 mcg/ml (mean, 80 to 81 mcg/ml) vs. Placebo for 8
wk | 2-wk washout of previous psychotropic medication
(6 wk for fluoxetine) | Diphenhydramine or hydroxyzine | 17-item HRSD, Hamilton Rating Scale for Anxiety
(HRSA), Clinical Global Impression (CGI), Clinician Administered Rating
Scale for Mania (CARS-M) at baseline then weekly; adverse events recorded
weekly; valproate serum concentrations and liver function tests at 4 and 8
wk | Not reported by treatment group Age, mean 9range),
y: 41 (25 to 54) M / F: 89% /
11% Caucasian: 81% | Veterans; otherwise characteristics not
reported | Numbers screened and eligible not reported / 25
enrolled / 25 randomized | 13 withdrew / 0lost to follow-up / 25
analyzed | Divalproex (N = 13) vs. Placebo (N
= 12)
HRSD (Primary Efficacy Measure), mean
percentage change from baseline to 8 wk: −43.51 vs.
−27.00 (calculated difference, −16.51; p
= 0.0002)
HRSD, mean change from baseline to 8 wk
(estimated from Figure 1 in
original report): −11.5 vs. −6.8 (calculated
difference, −4.7; p = 0.0002) Mixed-effects
model repeated measures (MMRM) analysis of results over time were
significant in favor of divalproex (p=0 033) | HRSA, mean percentage change: −35.21
vs. −5.25; calculated difference, 29.96; p =
0.0001)
HRSA, mean change from baseline at wk 8 (estimated
from Figure 2 of original
report): −7 vs. −1.4 (calculated difference,
− 5.6) (p=0.033)
MMRM analysis of
results over time were significantly in favor of divalproex
(p=0.0001) | Rate of HRSD improvement (change over time using
random regression analysis), points improvement per time unit on square root
scale: 5.5 vs. 2.6 (calculated difference, 2.9; p =
0.0227)
Rate of HRSA improvement: 3.4 vs. 0.7 (calculated
difference, 2.7; p = 0.009) | CARS-M and CGI: NSD (data not reported) | Monitoring | Not reported | Divalproex vs. Placebo Total withdrawals: 6 / 13
(46.2%) vs. 7 / 12 (58.3%) Withdrawals due to AEs: 1
/ 13 (7.7%) vs. 0 / 12 (0.0%) | Most of the outpatient subjects were moderately
ill. This trial is unique for monitoring anxiolytic effects (which are not
typically evaluated in bipolar clinical trials). Results need to be
confirmed in larger, well- designed trials before one can conclude efficacy
of divalproex for acute treatment of bipolar depression. |